Research Article Capacity Analysis and Optimization in...
Transcript of Research Article Capacity Analysis and Optimization in...
Research ArticleCapacity Analysis and Optimization in Heterogeneous Networkwith Adaptive Cell Range Control
Xinyu Gu1 Xin Deng1 Qi Li1 Lin Zhang1 and Wenyu Li2
1 Beijing University of Posts and Telecommunications Beijing 100876 China2 Beijing Key Lab of New Generation Broadband Wireless Mobile Communication Technology Standard and VerificationChina Academy of Telecommunication Research Beijing 100191 China
Correspondence should be addressed to Xinyu Gu guxinyubjgmailcom
Received 20 February 2014 Revised 7 April 2014 Accepted 8 April 2014 Published 29 April 2014
Academic Editor Xiang Zhang
Copyright copy 2014 Xinyu Gu et alThis is an open access article distributed under theCreativeCommonsAttribution License whichpermits unrestricted use distribution and reproduction in any medium provided the original work is properly cited
As an attractive means of expanding mobile network capacity heterogeneous network is regarded as an important direction ofmobile network evolution To increase the capacity of for example hot spots a typical scenario in heterogeneous network is thatthe coverage areas of low power nodes (LPNs) are overlapped with macrocell To increase the utilization of small cells generatedby LPNs cell range extension (CRE) is used to extend the coverage of the small cells by adding cell specific offset (CSO) to smallcells during cell selection procedure The value of CSO however needs to be set carefully In this paper the capacity of users inmacrocells users in small cells and users in range extension areas is analyzed thoroughly in conditionswith andwithoutCRE Basedon the analysis an adaptive CSO updating algorithm is proposed The proposed algorithm updates the CSO value periodicallyby predicting the overall capacity and a new CSO value is selected which can give the optimal overall capacity The proposedalgorithm is evaluated by system-level simulations Simulation results indicate that the proposed algorithm can ensure a nearlyoptimal performance in all tested traffic load situations
1 Introduction
A heterogeneous network is typically composed of multipleradio access technologies transmission solutions and basestations of varying transmission power [1] To meet theincreasing traffic demands and fulfilling usersrsquo high expec-tations for mobile broadband heterogeneous network hasbeen included in the evolution version of long-term evolu-tion (LTE) radio access technology named LTE-advancedin the third generation partnership project (3GPP) [2 3]togetherwith higher ordermultiple input andmultiple output(MIMO) [4] carrier aggregation (CA) [5] and coordinatedmultipoint transmission and reception (CoMP) [6]
In a heterogeneous network one typical scenario is thatthe coverage area of the low power nodes (LPN) and macro-cells are overlapped with each other Considering the casethat themacrocellrsquos carrier frequencies are reused throughoutthe network the interference between macrocells and LPNsis a very essential issue which needs to be considered in
order to optimize the network performance and end-userexperience especially with cell range extension (CRE) ofLPNs Poorly planned LPNs location or configuration notonly results in unsatisfactory experience in its coverage areabut also may cause performance degradation and overloadstate in macrocell [7] CRE was proposed in 3GPP in Release10 [8] with the purpose to enlarge the coverage area of LPNsand increase the served users in LPNs so as to explore thesystem resource (eg power spectrum) more efficiently
CRE is obtained by setting a cell selection offset (CSO)to LPNs during cell selectionreselection procedure so thatthe end-user would select the LPN with the higher priorityHowever when extending the served area of LPNs the usersthat locate in the extended range could suffer from severeinterference from themacro base station due to the differencebetween the transmit power of the macro base stations andthe LPNs This is illustrated by Figure 1
As discussed above there is a trade-off between the LPNoffloading ratio and the user experienced interference in
Hindawi Publishing CorporationInternational Journal of Antennas and PropagationVolume 2014 Article ID 215803 10 pageshttpdxdoiorg1011552014215803
2 International Journal of Antennas and Propagation
Signal from macrocell is the strongest
Signal from LPNis the strongest
A cell selection offsetcan be used to extend
the rangeSignal from macrocell is the strongest in the extended range as the macro base
station has higher transmit power
Figure 1 Illustration of cell range extension
LPNs especially in the extended LPN rangeTherefore how toproperly set CSO is very necessary to be studied In previousinvestigations CSOwas usually decided through simulationsfor example in [9] and some fixed value was proposed in3GPP based on system-level simulations However simula-tions cannot cover the varying situations in realityThereforeit is necessary to discuss the optimal CSO setting fromtheoretical point of view In [10] a semianalytic tool for inves-tigating the capacity and fairness of heterogeneous networkswith range extension was provided With this tool the sumcapacity and other metrics of interest can be evaluated as acontinuous function of the CRE It is an important attemptto optimize CSO setting through an analytical way than inpure simulation However there are still some limitationsin [10] Firstly in the analysis the relationship between thecell load and the signal to interference and noise ratio(SINR) was not considered Secondly the load estimationof each cell was simply through the number of users ineach cell but the number of users may not be the mostsuitable metric to directly represent the actual cell loadFor example the traffic volume of each user also plays animportant role in actual cell load In [11] spectral efficiencywas analyzed for heterogeneous network with CRE andthe main conclusion was that the CRE biasing deterioratesthe outage and rate of the overall network by lowering theSINR if assuming full queues at all base stations Howeverthis assumption may not be true in reality and needs to berelaxed as pointed out by [11] Besides theoretical analysisconsidering the varying conditions in the real wireless net-work self-organizing network (SON) is a powerful tool toadaptively adjust the network configuration parameters torealize self-optimizing [12ndash14] In scenario of heterogeneousnetwork SON is still an effective way to achieve networkself-configuration and self-optimizing [15] There were someattempts to use SON to adaptively adjust the CSO settingfor example in [16 17] In [16] an adaptive algorithm todecide the CSO value was proposed based on the end-userperformance feedback However there are limitations withthe algorithm in [16] Firstly the number of users was usedto decide the load in each cell which is not accurate enoughin reality as pointed out Secondly it was proposed to usecell edge user throughput and average user throughput asinputs to the adaptation algorithm However both averageuser throughput and cell edge user throughput cannot beobtained from the network side directly It requires feedback
from users and increases uplink signaling overhead In [17]the authors proposed an adaptive CRE controlling techniquethat improves the cell edge user throughput in heterogeneousnetwork in which UE can automatically choose an optimalCSO from either CSOhigh or CSOlow However the setting ofCSOhigh and CSOlow was not sufficiently discussed
In this paper the capacity of users in macrocells usersin small cells and especially users in range extension areas isanalyzed thoroughly in conditions with and without CRE Inthe analysis cell load is estimated based on the resource block(RB) utilization ratio Furthermore the relationship betweencell load and SINR is considered in the analysis Conditions toachieve higher capacity by CRE are derived respectively forusers served bymacrocell and LPNs Based on the analysis anadaptive CSO updating algorithm is proposedThe proposedalgorithm updates the CSO value periodically by predictingthe overall capacity and a new CSO value is selected whichcan give the optimal overall capacity
2 Capacity Analysis with Cell Range Extension
In this section the capacities of users in macrocells and insmall cells are analyzed especially for users in CRE regionUsers in CRE region refer to users whose serving cell changeswith CRE for example served by macrocell without CRE butserved by LPN with CRE
In this paper a mathematical framework for quantitativeinvestigation of self-optimizing wireless network (SON) [1819] is partly reused for the basis of the analysis
(i) A network layout by network nodes eithermacrocellsor LPNs defining a cell 119888 at the position 119902
119888
(ii) user equipment (UE) 119906 located at position 119902119906 in
dynamic simulation the user positions can changeover simulation time
(iii) UE 119906 is served by cell 119888 = 119878(119906) where 119878(119906) is theconnection function and every user is connectedexactly to a single base station (BS)
(iv) a pathloss mapping 119871119888( 119902119906) defined by the positions
of user 119906 relative to a cell 119888 the pathlossmapping takesall position-dependent channel model effects intoaccount for example distance-dependent pathlossshadowing and angle-dependent antenna patterns
(v) a cell load 120588119888 which defines the ratio of used resources
in LTE physical resource blocks (PRBs) to all availableresources in a given cell
(vi) a transmit power 119875119888for cell 119888 and thermal noise119873
Based on the above definitions the SINR of user 119906denoted by 120574
119906 with serving cell 119878(119906) can be written as
120574119906=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(1)
21Mechanism forCell Range Extension Typically the servedarea of a LPN is smaller than the served area of a macro basestation This is in line with the criterion of cell selection For
International Journal of Antennas and Propagation 3
cell selection usually the UE measures the reference signalreceived power (RSRP) from each cell and selects the servingcell with the maximum received signal power as shown by
119878 (119906) = argmax119888
(RSRP119906119888) = argmax
119888
(119875119888sdot 119871119888( 119902119906)) (2)
In (2) UE119906 selects cell c as its serving cell 119878(119906) because theRSRP from cell 119888 is the maximum among the measurementsfrom different cells
In homogeneous network the cell with the maximumRSRP is usually the cell with the best path-gain to thisUE however in heterogeneous network where the deployedbase stations have different transmit powers the cell withmaximum RSRP may no longer be the cell with the bestpath-gain to the UE As shown in (2) the RSRP of user 119906received from cell 119888 is the transmit power of cell 119888multipliedby path-gain (119871
119888) between cell 119888 and user 119906 (at the position
119902119906) Considering a typical transmit power of macrocell and
picocell as 43 dBm and 30 dBm respectively only UE whosepath-gain to the picocell is 13 dB better than the path-gain tomacrocell can select the picocell as its serving cell This leadsto a small number of served users in LPNs
To cope with this issue CRE is proposed for heteroge-neous network in 3GPPWithCRE a cell specific bias CSO isadded to the measured RSRP during cell selection procedureas shown in
119878 (119906) = argmax119888
(RSRP119906119888+ CSO119888) (3)
TheCSO value is usually zero formacrocell and a positivevalue for a LPN
22 Capacity Analysis with Cell Range Extension Assumingideal link adaptation which means that the most suitablemodulation and coding scheme (MCS) is selected accordingto the SINR of user 119906 the achievable data rate can berepresented by the Shannon capacity as
119877 (120574119906) = log
2(1 + 120574
119906) (4)
Here 119877(120574119906) is the achievable data rate if user 119906 is sched-
uled Considering the normal case that more than one user isconnecting to the same cell and sharing the resource in thesame cell and by assuming a round robin scheduling strategythe achievable capacity can be represented by
119862 (119906) = (1 minus 120588119878(119906)) log2(1 + 120574
119906) (5)
where120588119878(119906)
is the load in the serving cell 119878(119906) Intuitively usersin a cell with high load (large 120588
119888) have lower probability to get
chance to be scheduled and even if the SINR of this user isgood the final capacity is still not high
221 Capacity Analysis for Users in CRE Region In this sec-tion capacity of users in the CRE region is analyzed For theseusers before a CSO value is applied the serving cell of theseusers is the macrocell and after the CSO value is applied theserving cell of these users is the LPNWe analyze the capacitychange of these users with and without CRE
Denote the capacity of user 119906 in CRE region by 119862CRE(119906)For these users the serving cell 119878(119906) will change with andwithout CRE Represent the serving cell before applying CSOas 119878(119898)(119906) and the serving cell after applying CSO as 119878(119901)(119906)The SINR of user 119906 before applying CSO is 120574(119898)
119906CRE
120574(119898)
119906CRE =119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906)
119873 + sum119888 = 119878(119898)(119906)120588119888119875119888119871119888( 119902119906)
(6)
To facilitate further analysis the interference from thepicocell which is the serving cell after applying CSO isseparated from the overall interference Therefore the SINRof user 119906 before applying CSO is
120574(119898)
119906CRE =119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906)
119873 + sum119888 = 119878(119898)(119906)120588119888119875119888119871119888( 119902119906)
= (119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906))
times(119873 + sum
119888 = 119878(119898)(119906)
119888 = 119878(119901)(119906)
120588119888119875119888119871119888( 119902119906)
+ 120588119878(119901)(119906)119875119878(119901)(119906)119871119878(119901)(119906)( 119902119906))
minus1
(7)
Suppose the SINR of user 119906 after applying CSO is 120574(119901)119906CRE
120574(119901)
119906CRE =119875119883(119901)(119906)119871119883(119901)(119906)( 119902119906)
119873 + sum119888 =119883(119901)(119906)120588119888119875119888119871119888( 119902119906)
(8)
With CRE the received power from the previous servingcell becomes the interference Similarly the interference fromthe macrocell which is the serving cell before applying CSOis extracted from the overall interference for analysis TheSINR of user u after applying CSO could be written as
120574(119901)
119906CRE =119875119878(119901)(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
= (119875119878(119901)(119906)119871119878(119901)(119906)( 119902119906))
times(119873 + sum
119888 = 119878(119898)(119906)
119888 = 119878(119901)(119906)
120588119888119875119888119871119888( 119902119906)
+120588119878(119898)(119906)119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906))
minus1
(9)
4 International Journal of Antennas and Propagation
Table 1 System evaluation assumptions and parameters
Parameters Valuesdescriptions
Macro base stationdeployment
19 3-sector cells with wrap-aroundSite to site distance 500mHeight 40m above groundTx power 46 dBm
LPN deployment
6 LPNs per macrocell on averageHeight 5m above groundTx power 30 dBmCell radius 30m
Propagation model ITU UMa [20]
Antenna model Macro 3GPP [21]LPN dipole
Transmission scheme SU-MIMO 2Tx 2RxScheduler Proportional fairTraffic type DownloadNumber of users permacrocell 75
User distribution Uniform or in hotspot
For a user in the cell range extension area although theselected serving cell is the picocell the received power frommacro base station is still larger than received signal powerfrom picocell This means that for these users it is true that
119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906) ge 119875119878(119901)(119906)119871119878(119901)(119906)( 119902119906) (10)
Considering the most common case that the load in themacrocell is not lower than that in the picocell comparing(7) and (9) it can be concluded that the UE who are ldquoforcedrdquointo picocells by range extension will suffer from lower SINRThat is
120574(119901)
119906CRE le 120574(119898)
119906CRE (11)
For example Figure 2 clearly shows the SINR loss of usersin the cell range extension area In Figure 2 the cell selectionoffset is 6 dB and more simulation settings can be found inTable 1
Based on the analysis of the change of SINRof users in cellrange extension region the capacity of these users with andwithout cell range extension is further analyzed Suppose thecapacity of user 119906 before and after applying CSO is 119862(119898)CRE(119906)
and 119862(119901)CRE(119906) respectively
119862(119898)
CRE (119906) = (1 minus 120588119878(119898)(119906)) log2 (1 + 120574(119898)
119906CRE) (12a)
119862(119901)
CRE (119906) = (1 minus 120588119878(119901)(119906)) log2 (1 + 120574(119901)
119906CRE) (12b)
Comparing (12a) and (12b) it can be seen that althoughthe SINR is expected to be reduced (120574(119901)
119906CRE le 120574(119898)
119906CRE) forusers in CRE region after range extension the capacity ofthese users depends on the cell load of their serving cells aswell If the cell load in the picocell is below the cell load in
0 5 10 15 20 250
10
20
30
40
50
60
70
80
90
100
CDF
()
SINR of users in CRE region
SINR (dB)
wo CREw CRE
minus15 minus10 minus5
Figure 2 SINR of users in range extension area
the macrocell by certain extent although there is an expectedSINR loss the capacity of users in the CRE region can stillexpect an increase
119862(119901)
CRE (119906) gt 119862(119898)
CRE (119906) (13a)
(1minus120588119878(119901)(119906)) log2(1+120574(119901)
119906CRE) gt (1minus120588119878(119898)(119906)) log2 (1+120574(119898)
119906CRE)
(13b)
Therefore we can derive the theorem below to show thecondition that ensures a gain in capacity of users in CREregion
Theorem 1 The capacity of users in region of CRE can beincreased if and only if the cell load in the original servingmacrocell before CRE satisfies the inequality as below
120588119878(119898)(119906)gt 1 minus (1 minus 120588
119878(119901)(119906))
log2(1 + 120574
(119901)
119906CRE)
log2(1 + 120574
(119898)
119906CRE) (14)
Comparing the role of SINR and cell load in capacityit can be seen that the cell load linearly contributes tothe capacity while SINR contributes in logarithmic scaleTherefore the contribution from cell load plays an evenmoreimportant role in capacity than SINR does and needs to betreated seriously
To summarize on one hand for users in the CREregion usually the SINR decreases due to the suboptimalreceived signal power and high interference from macrocellon the other hand CRE may also bring benefit in terms ofavailable resource (eg frequency time and power) to theseusers More resource to a user means for example moreopportunity to be scheduled by the networkwhich eventuallycontributes to the user capacity
International Journal of Antennas and Propagation 5
Suppose the total capacity of users in CRE region is119862CREand the number of users in the CRE region is119873
119877 we have
119862CRE =
119873119877
sum
119906=1
119862CRE (119906) (15)
222 Capacity Analysis for Users inMacrocell In this sectionthe capacity of users who are served by the macrocellconsistently before and after range extension is analyzed Forthese users although the serving cell is unchanged the SINRand capacity are impacted anyway by range extension
Suppose the SINR of a user 119906 in this case is representedby 120574119906119872
120574119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(16)
By applying range extension more users are connectedto the picocell and the cell load in the picocell is expectedto increase To facilitate the analysis the interference frompicocell is separated from the total interference in (16)Denoting the picocell as119901 the SINRof user119906 can be rewrittenas
120574119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 = 119901
120588119888119875119888119871119888( 119902119906) + 120588119901119875119901119871119901( 119902119906)
(17)
Denote the SINR of user 119906 after CRE as 1205741015840119906119872
Afterapplying CRE the received signal power from the servingmacrocell can be assumed to be unchanged if we assumethe userrsquos position does not change significantly during thetime of handover executionThe interference from other cellsexcept for the picocell can also be assumed to be unchangedfor simplicity However with CRE the load in the picocell1205881015840
119901 is expected to increase (1205881015840
119901gt 120588119901) Consequently the
interference from picocell will increase eventually The SINRof user 119906 after CRE is then
1205741015840
119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 = 119901
120588119888119875119888119871119888( 119902119906) + 1205881015840
119901119875119901119871119901( 119902119906)
(18)
Therefore the SINR of users served by the macrocell willdecrease due to the increased interference from picocell
1205741015840
119906119872lt 120574119906119872 (19)
Denoting the macrocell as119898 and the capacity of users inmacrocell as119862
119872(119906) the capacity is determined by both SINR
of these users and the load situation in the macrocell
119862119872(119906) = (1 minus 120588
119898) log2(1 + 120574
119906119872) (20)
Denoting the capacity of user 119906 after CRE as 1198621015840119872(119906) we
have
1198621015840
119872(119906) = (1 minus 120588
1015840
119898) log2(1 + 120574
1015840
119906119872) (21)
It is desirable that with CRE the capacity of users stay inmacrocell could increase
1198621015840
119872(119906) gt 119862
119872(119906) (22a)
(1 minus 1205881015840
119898) log2(1 + 120574
1015840
119906119872) gt (1 minus 120588
119898) log2(1 + 120574
119906119872) (22b)
Therefore we can derive the theorem as below to show thecondition that ensures a gain in capacity with CRE for usersstay in macrocell
Theorem2 Thecapacity of users inmacrocell can be increasedif and only if the cell load in themacrocell after CRE satisfies theinequality as follows
1205881015840
119898lt 1 minus (1 minus 120588
119898)
log2(1 + 120574
119906119872)
log2(1 + 120574
1015840
119906119872)
(23)
Theorem 2 shows that only when the cell load of macro-cell decreases low enough to compensate for the loss in SINRthe capacity of these users can increase
More specifically suppose the number of users in themacrocell is 119873
119872 the number of scheduled resource blocks
for each user 119906 is NRB119906 and the total number of available
resource blocks in the cell is NRBavail119898 then the macrocellload 120588
119898is
120588119898=
sum119873119872
119906=1NRB119906
NRBavail119898 (24)
Therefore the capacity of user 119906 is
119862119872(119906) = (1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872) (25)
The number of scheduled resource blocks is usually afunction of SINR Considering a typical traffic type that isnamed as constant bit rate (CBR) service as the example
NRB119906=
119863119906
119877 (120574119906119872) sdot BW (26)
where BW is the bandwidth of one PRB (eg in LTE systemBW is 180 kHz) and 119863
119906is the demanded data rate of user 119906
As shown in (4) 119877(120574) is the achievable data rate that can beestimated according to the Shannon capacity Therefore thecapacity of users stay in macrocell is actually a function ofboth SINRand the number of users inmacrocell Consideringthat the served user number linearly contributes to thecapacity while SINR contributes in logarithmic scale fromthe capacity of macrocell point of view it is desirable tohandover more users out of macrocell However from theoverall capacity point of view it is necessary to consider thecapacity change of both users stay in macrocell and usersstay in CRE region as well as users in picocell which we willanalyze in the following section
6 International Journal of Antennas and Propagation
Suppose the total capacity of users in macrocell is119862119872 we
have
119862119872=
119873119872
sum
119906=1
119862119872(119906)
=
119873119872
sum
119906=1
(1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872)
(27)
223 Capacity Analysis for Users in Picocell The capacity ofusers served by picocell both before and after applying CRE isanalyzed similarly For these users although the serving cellis unchanged the SINR and capacity are eventually impactedby range extension
Suppose the SINR of a user 119906 in this case is presented by120574119906119875
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(28)
To facilitate the analysis the macrocell interference isseparated from the total interference in (28) Denoting themacrocell as119898 the SINR of user 119906 can be rewritten as
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 120588119898119875119898119871119898( 119902119906)
(29)
Denote the SINR of user 119906 after CRE as 1205741015840119906119875
Similarto the analysis for users in macrocell after applying CREthe received signal power from the serving picocell can beassumed to be unchanged and the interference from othercells except for the macrocell can also be assumed to beunchanged for simplicity However with CRE the load in themacrocell denoted by 1205881015840
119898 is changed and usually the load in
macrocell is expected to be reduced (1205881015840119898lt 120588119898) Consequently
the interference from macrocell will decrease The SINR ofpico user u after CRE is then
1205741015840
119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 1205881015840
119898119875119898119871119898( 119902119906)
(30)
Therefore the SINR of users served by picocell increasesdue to the reduced interference as
1205741015840
119906119875gt 120574119906119875 (31)
Denoting the picocell as 119901 and the capacity of user119906 in picocell with and without CRE as 119862
119901(119906) and 1198621015840
119901(119906)
respectively then
119862119875(119906) = (1 minus 120588
119875) log2(1 + 120574
119906119875) (32)
1198621015840
119875(119906) = (1 minus 120588
1015840
119875) log2(1 + 120574
1015840
119906119875) (33)
To achieve capacity gain for users in picocell that is
1198621015840
119875(119906) gt 119862
119875(119906) (34a)
(1 minus 1205881015840
119875) log2(1 + 120574
1015840
119906119875) gt (1 minus 120588
119875) log2(1 + 120574
119906119875) (34b)
Therefore we derive the following theorem which statesthe conditions for users in picocell to achieve capacity gain byCRE
Theorem 3 The capacity of users in picocell can be increasedif and only if the cell load in the picocell after CRE satisfies theinequality as follows
1205881015840
119901lt 1 minus (1 minus 120588
119875)
log2(1 + 120574
119906119875)
log2(1 + 120574
1015840
119906119875)
(35)
FromTheorem 3 it can be concluded that if the increasedcell load in picocell does not exceed the limit as shown ininequality (35) the capacity of users in picocell can increaseOtherwise the capacity of the pico users would decrease
Suppose the number of users in the picocell is 119873119875 the
number of scheduled resource blocks for each user 119906 is NRB119906
and the total number of available resource blocks in thepicocell is NRBavail119901 The picocell load 120588
119901is then
120588119901=
sum119873119875
119906=1NRB119906
NRBavail119901 (36)
And the picocell load after CRE is
1205881015840
119901=
sum119873119875+119873119877
119906=1NRB1015840119906
NRBavail119901 (37)
where 119873119877is the number of users in CRE region and NRB1015840
119906
is the number of resource blocks of user 119906 served by pico Asshown in (26) the number of resource blocks of user119906usuallydepends on its SINR According to the analysis typically theSINR of users in CRE region would decrease while the SINRof users in picocell would increase which results in eitherincreased NRB or decreased NRB However considering theincreased number of served users119873
119877 the cell load in picocell
usually gets difficult to satisfy the inequality of (35) andresults in a decreased capacity
Suppose the total capacity of users in picocell is 119862119875 we
have
119862119875=
119873119875
sum
119906=1
119862119875(119906)
=
119873119875
sum
119906=1
(1 minus
sum119873119875
119906=1NRB119906
NRBavail119875) log2(1 + 120574
119906119875)
(38)
224 Overall Capacity Summarizing capacity of users inmacrocell in picocell and in the CRE region we can get theoverall capacity as
119862total = 119862119872 + 119862119875 + 119862CRE (39)
International Journal of Antennas and Propagation 7
From the analysis above it can be observed that withCRE the overall capacity not only depends on the change ofSINRs of users in macrocell in picocell and in CRE regionbut also depends on the load situation and change in bothmacrocell and picocell CRE may not always give benefit tothe overall capacity in the system Considering an examplethat if the traffic load in macrocell is low and there areenough radio resources that can be allocated to the servedusers the throughput of these users is consequently limited bytheir channel conditions that is SINRs In this case forcingthese users into the picocell by CRE can further damage theSINRs and consequently damage the usersrsquo throughput Onthe other hand if the traffic load in macrocell is high and theusersrsquo throughputs are limited by the available radio resourcehanding over these users to LPN by CRE can improve theusersrsquo throughputs
To get the optimal overall capacity in the next section anadaptive algorithm to optimize theCSO setting via predictionof overall capacity for different CSO values is proposed
3 Capacity Optimization Algorithm withAdaptive Cell Range Control
In this section an adaptive algorithm is proposed to optimizethe overall capacity by adjusting the CSO setting In thisalgorithm the CSO setting is updated periodically At eachupdate instance the current CSO value is updated by a newvalue if the predicted overall capacity can be increased by thisnew attempting valueThe attemptingCSO setting is obtainedby increasing or decreasing a step to the current CSO Thenew CSO value which gives the best overall capacity is thenselected The algorithm is described step by step as in below
31 Step 1 User Grouping with Attempting CSO SettingsSuppose CSO updating step is ΔCSO At each update instance119905 CSO value is updated based on current CSO value CSO
119905minus1
whichwas set at time interval 119905minus1The attemptingCSO valueCSO119905att is obtained by either increasing or decreasingCSO119905minus1
by the updating step ΔCSO as
CSO119905att = CSO119905inc = CSO119905minus1 + ΔCSO (40a)
CSO119905att = CSO119905dec = CSO119905minus1 minus ΔCSO (40b)
Based on the RSRP measurements from both macrocelland picocell for the attempted CSO setting all of the userscan be divided into three groups
Group 1Macro users These users stay in the macrocell bothwith the previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att le RSRP119898
119906
(41)
Group 2 Pico usersThese users stay in the picocell both withthe previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1gt RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(42)
Group 3 Users in CRE region These users will change theirserving cell with new CSO setting Take the case that CSO isincreased as the following example
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(43)
On the basis of the user grouping the capacities of usersin different groups are predicted respectively and the overallcapacity is predicted consequently
32 Step 2 Capacity Prediction with Attempting CSO SettingsIn step 2 the capacities of users in different groups arepredicted respectively according to the predicted SINR andcell load More details can be found below
321 Step 21 Capacity Prediction for Users in CRE RegionTo estimate the new capacity of users in CRE region we needto estimate the new SINR and predict the new cell load inpicocell and in macrocell
Taking the case that the CSO is increased by a step asexample to estimate the new SINR in picocell 120574(119901)
119906CRE we canuse
120574(119901)
119906CRE =RSRP(119901)
(RSRP(119898)120574(119898)119906CRE) + RSRP
(119898)minus RSRP(119901)
(44)
To estimate the new picocell load we need to estimate therequired number of RBs according to the new SINR For CBRservice this can be estimated similarly to (26) by
119873119906=
119863119906
119877 (120574(119901)
119906CRE) sdot BW (45)
For services other than CBR service the new requirednumber of RBs can be estimated according to the relative rela-tionship between the previous SINR in macrocell previousrequired number of RBs inmacrocell and the predicted SINRin picocell
119873119906
new= 119873119906
oldsdot
119877 (120574(119898)
119906CRE)
119877 (120574(119901)
119906CRE) (46)
Based on the prediction of the required number of RBsthe new picocell load can be estimated by assuming therequired number of RBs is unchanged for other users inpicocell
120588119901
new=
sum119873119875
119906=1119873119906+ sum119873119877
119906=1119873119906
new
NRBavail119901
(47)
8 International Journal of Antennas and Propagation
Consequently the new capacity of users newly added toCRE region can be estimated according to the analysis inSection 2
322 Step 22 Capacity Prediction for Macro Users and PicoUsers Formacro users to predict the SINR of users with newCSO the picocell load with new CSO needs to be estimatedThis can be achieved by method described in Step 21
The new cell load in macrocell can be simply estimatedby the new number of served users and assuming that therequired numbers of RBs for macro users are unchanged
The capacity of users in picocell can be estimated simi-larly
33 Optimal CSO Selection and Updating The predictednew capacities of users in different groups are summedtogether to get the overall predicted capacity The predictedoverall capacity is then compared with current capacity If theestimated new capacity is larger than the current capacity theCSO is updated
If 119862 (CSO119905att) gt 119862 (CSO119905minus1) CSO119905 = CSO119905att (48)
Otherwise CSO keeps unchangedIf the CSO is decreased by a step the overall capacity
can be estimated similarly by reversing the calculation shownabove If the new CSO value gives better predicted perfor-mance the CSO value is updated
With the proposed adaptive algorithm the CSO value isadjusted according to the predicted overall capacity in bothmacrocell and picocell By selecting the most suitable CSOvalue the overall capacity is expected to be optimized
4 Numerical Results
The proposed adaptive CSO updating algorithm is evaluatedvia system-level simulations In this section the simulationresults are presented and discussed The simulation assump-tions and parameter settings are illustrated in Table 1
Two typical scenarios are tested in the evaluation Inscenario 1 the users are uniformly distributed in the entiresimulated area LPNs are deployed randomly In scenario 2LPNs are deployed in hotspot areas and the hotspot areastake nearly 60of the usersThe base station deployment anduser distribution in two scenarios are illustrated in Figures 3and 4 To have a clear view only part of the simulated areas isillustrated
The percentages of users belonging to macrocell or LPNsin the two scenarios are illustrated in Figure 5 with differentCSO settings It can be observed that when CSO settingincreases the users belonging to LPNs increase It can beobserved as well that the percentages of users belonging toLPNs in hotspot scenario are much higher than in uniformscenario
Figure 6 shows the change of average cell load withdifferent CSO settings in different scenarios In uniformscenario most of the users are in macrocells and the loadin the macrocell is consequently quite high the cell load inpicocell increases with the increase in the CSO setting In
0
1
2
3
4
5
6
911
15 17
Figure 3 Cell layout and user distribution in uniform scenario
0
1
2
3
4
5
6
911
15 17
Figure 4 Cell layout and user distribution in hotspot scenario
hotspot scenario the average cell load inmacrocells decreaseswith the increase in CSO setting while the average cell load inLPNs increases further with the increased CSO setting
In Figure 7 the average sector throughput gain withdifferent CSO settings is compared with the proposed adap-tive algorithm for uniform scenario The average sectorthroughput with CSO = 0 dB is used as reference to obtainthe average sector throughput gain It can be observed that inthis scenario a higherCSO setting is desirable to offloadmoreusers to LPNs The proposed adaptive algorithm can achievethe similar throughput to the throughput of the optimal CSOsetting
In Figure 8 the average sector throughput gain with dif-ferent CSO settings is compared with the proposed adaptivealgorithm in hotspot scenarioThe average sector throughput
International Journal of Antennas and Propagation 9
0001020304050607080910
Uniform-macroUniform-pico
Hotspot-picoHotspot-macro
CSO 15CSO 10CSO 5CSO 0
Con
nect
ion
ratio
Figure 5 Connection ratio with different CSOs
0 2 4 6 8 10
05
06
07
08
09
10
Uniform-macroUniform-pico
Hotspot-macroHotspot-pico
CSO (dB)
Nor
mal
ized
cell
load
Figure 6 Cell load with different CSOs
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 7 Average sector throughput gain in uniform scenario
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 8 Average sector throughput gain in hotspot scenario
with CSO = 0 dB is used as reference to obtain the averagesector throughput gain In this scenario CSO of 10 dB isdesirable to balance the offloading gain and SINR loss Theproposed algorithm achieves the similar performance to theoptimal CSO setting
In different scenarios the optimal CSO setting is differ-ent However the proposed adaptive algorithm can achievesimilar performance with the optimal CSO setting in bothscenarios
5 Conclusion
As an attractivemeans of expandingmobile network capacityheterogeneous network has been included in LTE-advancedCell range extension is an important feature in heterogeneousnetwork to improve the utilization of the resource of lowpower nodes However the users in the cell range extensionarea usually suffer from high interference from macrocellsTo avoid the unnecessary interference it is desirable tocarefully consider the optimal cell selection offset value Inthis paper the capacity of users in macrocell and small cellsis analyzed thoroughly The conditions to improve capacityvia cell range extension are derived respectively for usersboth in macrocell and in small cell Based on the analysis anadaptive small cell coverage control algorithm is proposedWith the proposed adaptive algorithm the cell selectionoffset is updated periodically based on the prediction of theoverall system capacityThe algorithm is evaluated by system-level simulations and the results indicate that with theproposed algorithm a nearly optimal system performancecan be achieved in all tested load cases
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
10 International Journal of Antennas and Propagation
Acknowledgments
This work was supported by the State Major Science andTechnology Special Projects (Grant no 2013ZX03001026-001) and the Fundamental Research Funds for the CentralUniversities (Grant no 2014RC0107)
References
[1] A Damnjanovic J Montojo Y Wei et al ldquoA survey on 3GPPheterogeneous networksrdquo IEEE Wireless Communications vol18 no 3 pp 10ndash21 2011
[2] E Dahlman S Parkvall and J Skold 4G LTELTE-Advanced forMobile Broadband Academic Press 2011
[3] A Ghosh R Ratasuk B Mondal N Mangalvedhe and TThomas ldquoLTE-advanced next-generation wireless broadbandtechnologyrdquo IEEE Wireless Communications vol 17 no 3 pp10ndash22 2010
[4] X Zhang XGuW Li L Zhang J Shen andYWan ldquoThe studyof indoor and field trials on 2times8MIMOarchitecture in TD-LTEnetworkrdquo International Journal of Antennas and Propagationsvol 2013 Article ID 181579 9 pages 2013
[5] G Yuan X Zhang WWang and Y Yang ldquoCarrier aggregationfor LTE-advancedmobile communication systemsrdquo IEEE Com-munications Magazine vol 48 no 2 pp 88ndash93 2010
[6] R Irmer H Droste P Marsch et al ldquoCoordinated multipointconcepts performance and field trial resultsrdquo IEEE Communi-cations Magazine vol 49 no 2 pp 102ndash111 2011
[7] I Siomina and Y Di ldquoLoad balancing in heterogeneous LTErange optimization via cell offset and load-coupling character-izationrdquo in Proceedings of the IEEE International Conference onCommunications pp 1357ndash1361 2012
[8] The 3rd Generation Partnership Project (3GPP) ldquoFeasibilitystudy for further advancements for E-UTRA (LTEAdvanced)(Release 10)rdquo Tech Rep TR 36912 2011
[9] The 3rd Generation Partnership Project(3GPP) ldquoSystem per-formance of heterogeneous networks with range expansionrdquoTech Rep R1-101203 Samsung 2010
[10] I Guvenc ldquoCapacity and fairness analysis of heterogeneous net-works with range expansion and interference coordinationrdquoIEEECommunications Letters vol 15 no 10 pp 1084ndash1087 2011
[11] H-S Jo Y J Sang P Xia and J G Andrews ldquoHeterogeneouscellular networks with flexible cell association a comprehensivedownlink SINR analysisrdquo IEEE Transactions on Wireless Com-munications vol 11 no 10 pp 3484ndash3495 2012
[12] ldquoSelf-optimisationand self-configuration in wireless networksrdquoSOCRATES European Research Project httpwwwfp7-socratesorg
[13] The 3rd Generation Partnership Project(3GPP) ldquoSelf-config-uring and self-optimizing network use cases and solutionsrdquoTech Rep TR 36902 2009
[14] ldquoUse cases related to Self Organizing network Overall descrip-tionrdquoNext generationMobileNetworks httpwwwngmnorg
[15] X Chu and D Lopez-Perez Heterogeneous Cellular NetworksTheory Simulation and Deployment Cambridge UniversityPress 2013
[16] P Tian H Tian J Zhu L Chen and X She ldquoAn adaptive biasconfiguration strategy for range extension in LTE-Advancedheterogeneous networksrdquo inProceedings of the IET InternationalConference on Communication Technology and Application(ICCTA rsquo11) pp 336ndash340 2011
[17] K Kikuchi and H Otsuka ldquoProposal of adaptive control CREin heterogeneous networksrdquo in Proceedings of the IEEE Inter-national Symposium on Personal Indoor and Mobile RadioCommunications (PIMRC rsquo12) pp 910ndash914 2012
[18] I Viering M Dottling and A Lobinger ldquoA mathematical per-spective of self-optimizing wireless networksrdquo in Proceedings ofthe IEEE International Conference on Communications (ICCrsquo09) pp 1ndash6 June 2009
[19] A Lobinger S Stefanski T Jansen and I Balan ldquoLoad balanc-ing in downlink LTE self-optimizing networksrdquo in Proceedingsof the IEEE 71st Vehicular Technology Conference (VTC rsquo10) pp1ndash5 May 2010
[20] ldquoGuidelines for evaluation of radio transmission technologiesfor IMT-2000rdquo Recommendation ITU-R M1225 1997
[21] The 3rd Generation Partnership Project (3GPP) ldquoFurtheradvancements for E-UTRAN physical layer aspects (Release9)rdquo Tech Rep TS 36814 2010
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
2 International Journal of Antennas and Propagation
Signal from macrocell is the strongest
Signal from LPNis the strongest
A cell selection offsetcan be used to extend
the rangeSignal from macrocell is the strongest in the extended range as the macro base
station has higher transmit power
Figure 1 Illustration of cell range extension
LPNs especially in the extended LPN rangeTherefore how toproperly set CSO is very necessary to be studied In previousinvestigations CSOwas usually decided through simulationsfor example in [9] and some fixed value was proposed in3GPP based on system-level simulations However simula-tions cannot cover the varying situations in realityThereforeit is necessary to discuss the optimal CSO setting fromtheoretical point of view In [10] a semianalytic tool for inves-tigating the capacity and fairness of heterogeneous networkswith range extension was provided With this tool the sumcapacity and other metrics of interest can be evaluated as acontinuous function of the CRE It is an important attemptto optimize CSO setting through an analytical way than inpure simulation However there are still some limitationsin [10] Firstly in the analysis the relationship between thecell load and the signal to interference and noise ratio(SINR) was not considered Secondly the load estimationof each cell was simply through the number of users ineach cell but the number of users may not be the mostsuitable metric to directly represent the actual cell loadFor example the traffic volume of each user also plays animportant role in actual cell load In [11] spectral efficiencywas analyzed for heterogeneous network with CRE andthe main conclusion was that the CRE biasing deterioratesthe outage and rate of the overall network by lowering theSINR if assuming full queues at all base stations Howeverthis assumption may not be true in reality and needs to berelaxed as pointed out by [11] Besides theoretical analysisconsidering the varying conditions in the real wireless net-work self-organizing network (SON) is a powerful tool toadaptively adjust the network configuration parameters torealize self-optimizing [12ndash14] In scenario of heterogeneousnetwork SON is still an effective way to achieve networkself-configuration and self-optimizing [15] There were someattempts to use SON to adaptively adjust the CSO settingfor example in [16 17] In [16] an adaptive algorithm todecide the CSO value was proposed based on the end-userperformance feedback However there are limitations withthe algorithm in [16] Firstly the number of users was usedto decide the load in each cell which is not accurate enoughin reality as pointed out Secondly it was proposed to usecell edge user throughput and average user throughput asinputs to the adaptation algorithm However both averageuser throughput and cell edge user throughput cannot beobtained from the network side directly It requires feedback
from users and increases uplink signaling overhead In [17]the authors proposed an adaptive CRE controlling techniquethat improves the cell edge user throughput in heterogeneousnetwork in which UE can automatically choose an optimalCSO from either CSOhigh or CSOlow However the setting ofCSOhigh and CSOlow was not sufficiently discussed
In this paper the capacity of users in macrocells usersin small cells and especially users in range extension areas isanalyzed thoroughly in conditions with and without CRE Inthe analysis cell load is estimated based on the resource block(RB) utilization ratio Furthermore the relationship betweencell load and SINR is considered in the analysis Conditions toachieve higher capacity by CRE are derived respectively forusers served bymacrocell and LPNs Based on the analysis anadaptive CSO updating algorithm is proposedThe proposedalgorithm updates the CSO value periodically by predictingthe overall capacity and a new CSO value is selected whichcan give the optimal overall capacity
2 Capacity Analysis with Cell Range Extension
In this section the capacities of users in macrocells and insmall cells are analyzed especially for users in CRE regionUsers in CRE region refer to users whose serving cell changeswith CRE for example served by macrocell without CRE butserved by LPN with CRE
In this paper a mathematical framework for quantitativeinvestigation of self-optimizing wireless network (SON) [1819] is partly reused for the basis of the analysis
(i) A network layout by network nodes eithermacrocellsor LPNs defining a cell 119888 at the position 119902
119888
(ii) user equipment (UE) 119906 located at position 119902119906 in
dynamic simulation the user positions can changeover simulation time
(iii) UE 119906 is served by cell 119888 = 119878(119906) where 119878(119906) is theconnection function and every user is connectedexactly to a single base station (BS)
(iv) a pathloss mapping 119871119888( 119902119906) defined by the positions
of user 119906 relative to a cell 119888 the pathlossmapping takesall position-dependent channel model effects intoaccount for example distance-dependent pathlossshadowing and angle-dependent antenna patterns
(v) a cell load 120588119888 which defines the ratio of used resources
in LTE physical resource blocks (PRBs) to all availableresources in a given cell
(vi) a transmit power 119875119888for cell 119888 and thermal noise119873
Based on the above definitions the SINR of user 119906denoted by 120574
119906 with serving cell 119878(119906) can be written as
120574119906=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(1)
21Mechanism forCell Range Extension Typically the servedarea of a LPN is smaller than the served area of a macro basestation This is in line with the criterion of cell selection For
International Journal of Antennas and Propagation 3
cell selection usually the UE measures the reference signalreceived power (RSRP) from each cell and selects the servingcell with the maximum received signal power as shown by
119878 (119906) = argmax119888
(RSRP119906119888) = argmax
119888
(119875119888sdot 119871119888( 119902119906)) (2)
In (2) UE119906 selects cell c as its serving cell 119878(119906) because theRSRP from cell 119888 is the maximum among the measurementsfrom different cells
In homogeneous network the cell with the maximumRSRP is usually the cell with the best path-gain to thisUE however in heterogeneous network where the deployedbase stations have different transmit powers the cell withmaximum RSRP may no longer be the cell with the bestpath-gain to the UE As shown in (2) the RSRP of user 119906received from cell 119888 is the transmit power of cell 119888multipliedby path-gain (119871
119888) between cell 119888 and user 119906 (at the position
119902119906) Considering a typical transmit power of macrocell and
picocell as 43 dBm and 30 dBm respectively only UE whosepath-gain to the picocell is 13 dB better than the path-gain tomacrocell can select the picocell as its serving cell This leadsto a small number of served users in LPNs
To cope with this issue CRE is proposed for heteroge-neous network in 3GPPWithCRE a cell specific bias CSO isadded to the measured RSRP during cell selection procedureas shown in
119878 (119906) = argmax119888
(RSRP119906119888+ CSO119888) (3)
TheCSO value is usually zero formacrocell and a positivevalue for a LPN
22 Capacity Analysis with Cell Range Extension Assumingideal link adaptation which means that the most suitablemodulation and coding scheme (MCS) is selected accordingto the SINR of user 119906 the achievable data rate can berepresented by the Shannon capacity as
119877 (120574119906) = log
2(1 + 120574
119906) (4)
Here 119877(120574119906) is the achievable data rate if user 119906 is sched-
uled Considering the normal case that more than one user isconnecting to the same cell and sharing the resource in thesame cell and by assuming a round robin scheduling strategythe achievable capacity can be represented by
119862 (119906) = (1 minus 120588119878(119906)) log2(1 + 120574
119906) (5)
where120588119878(119906)
is the load in the serving cell 119878(119906) Intuitively usersin a cell with high load (large 120588
119888) have lower probability to get
chance to be scheduled and even if the SINR of this user isgood the final capacity is still not high
221 Capacity Analysis for Users in CRE Region In this sec-tion capacity of users in the CRE region is analyzed For theseusers before a CSO value is applied the serving cell of theseusers is the macrocell and after the CSO value is applied theserving cell of these users is the LPNWe analyze the capacitychange of these users with and without CRE
Denote the capacity of user 119906 in CRE region by 119862CRE(119906)For these users the serving cell 119878(119906) will change with andwithout CRE Represent the serving cell before applying CSOas 119878(119898)(119906) and the serving cell after applying CSO as 119878(119901)(119906)The SINR of user 119906 before applying CSO is 120574(119898)
119906CRE
120574(119898)
119906CRE =119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906)
119873 + sum119888 = 119878(119898)(119906)120588119888119875119888119871119888( 119902119906)
(6)
To facilitate further analysis the interference from thepicocell which is the serving cell after applying CSO isseparated from the overall interference Therefore the SINRof user 119906 before applying CSO is
120574(119898)
119906CRE =119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906)
119873 + sum119888 = 119878(119898)(119906)120588119888119875119888119871119888( 119902119906)
= (119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906))
times(119873 + sum
119888 = 119878(119898)(119906)
119888 = 119878(119901)(119906)
120588119888119875119888119871119888( 119902119906)
+ 120588119878(119901)(119906)119875119878(119901)(119906)119871119878(119901)(119906)( 119902119906))
minus1
(7)
Suppose the SINR of user 119906 after applying CSO is 120574(119901)119906CRE
120574(119901)
119906CRE =119875119883(119901)(119906)119871119883(119901)(119906)( 119902119906)
119873 + sum119888 =119883(119901)(119906)120588119888119875119888119871119888( 119902119906)
(8)
With CRE the received power from the previous servingcell becomes the interference Similarly the interference fromthe macrocell which is the serving cell before applying CSOis extracted from the overall interference for analysis TheSINR of user u after applying CSO could be written as
120574(119901)
119906CRE =119875119878(119901)(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
= (119875119878(119901)(119906)119871119878(119901)(119906)( 119902119906))
times(119873 + sum
119888 = 119878(119898)(119906)
119888 = 119878(119901)(119906)
120588119888119875119888119871119888( 119902119906)
+120588119878(119898)(119906)119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906))
minus1
(9)
4 International Journal of Antennas and Propagation
Table 1 System evaluation assumptions and parameters
Parameters Valuesdescriptions
Macro base stationdeployment
19 3-sector cells with wrap-aroundSite to site distance 500mHeight 40m above groundTx power 46 dBm
LPN deployment
6 LPNs per macrocell on averageHeight 5m above groundTx power 30 dBmCell radius 30m
Propagation model ITU UMa [20]
Antenna model Macro 3GPP [21]LPN dipole
Transmission scheme SU-MIMO 2Tx 2RxScheduler Proportional fairTraffic type DownloadNumber of users permacrocell 75
User distribution Uniform or in hotspot
For a user in the cell range extension area although theselected serving cell is the picocell the received power frommacro base station is still larger than received signal powerfrom picocell This means that for these users it is true that
119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906) ge 119875119878(119901)(119906)119871119878(119901)(119906)( 119902119906) (10)
Considering the most common case that the load in themacrocell is not lower than that in the picocell comparing(7) and (9) it can be concluded that the UE who are ldquoforcedrdquointo picocells by range extension will suffer from lower SINRThat is
120574(119901)
119906CRE le 120574(119898)
119906CRE (11)
For example Figure 2 clearly shows the SINR loss of usersin the cell range extension area In Figure 2 the cell selectionoffset is 6 dB and more simulation settings can be found inTable 1
Based on the analysis of the change of SINRof users in cellrange extension region the capacity of these users with andwithout cell range extension is further analyzed Suppose thecapacity of user 119906 before and after applying CSO is 119862(119898)CRE(119906)
and 119862(119901)CRE(119906) respectively
119862(119898)
CRE (119906) = (1 minus 120588119878(119898)(119906)) log2 (1 + 120574(119898)
119906CRE) (12a)
119862(119901)
CRE (119906) = (1 minus 120588119878(119901)(119906)) log2 (1 + 120574(119901)
119906CRE) (12b)
Comparing (12a) and (12b) it can be seen that althoughthe SINR is expected to be reduced (120574(119901)
119906CRE le 120574(119898)
119906CRE) forusers in CRE region after range extension the capacity ofthese users depends on the cell load of their serving cells aswell If the cell load in the picocell is below the cell load in
0 5 10 15 20 250
10
20
30
40
50
60
70
80
90
100
CDF
()
SINR of users in CRE region
SINR (dB)
wo CREw CRE
minus15 minus10 minus5
Figure 2 SINR of users in range extension area
the macrocell by certain extent although there is an expectedSINR loss the capacity of users in the CRE region can stillexpect an increase
119862(119901)
CRE (119906) gt 119862(119898)
CRE (119906) (13a)
(1minus120588119878(119901)(119906)) log2(1+120574(119901)
119906CRE) gt (1minus120588119878(119898)(119906)) log2 (1+120574(119898)
119906CRE)
(13b)
Therefore we can derive the theorem below to show thecondition that ensures a gain in capacity of users in CREregion
Theorem 1 The capacity of users in region of CRE can beincreased if and only if the cell load in the original servingmacrocell before CRE satisfies the inequality as below
120588119878(119898)(119906)gt 1 minus (1 minus 120588
119878(119901)(119906))
log2(1 + 120574
(119901)
119906CRE)
log2(1 + 120574
(119898)
119906CRE) (14)
Comparing the role of SINR and cell load in capacityit can be seen that the cell load linearly contributes tothe capacity while SINR contributes in logarithmic scaleTherefore the contribution from cell load plays an evenmoreimportant role in capacity than SINR does and needs to betreated seriously
To summarize on one hand for users in the CREregion usually the SINR decreases due to the suboptimalreceived signal power and high interference from macrocellon the other hand CRE may also bring benefit in terms ofavailable resource (eg frequency time and power) to theseusers More resource to a user means for example moreopportunity to be scheduled by the networkwhich eventuallycontributes to the user capacity
International Journal of Antennas and Propagation 5
Suppose the total capacity of users in CRE region is119862CREand the number of users in the CRE region is119873
119877 we have
119862CRE =
119873119877
sum
119906=1
119862CRE (119906) (15)
222 Capacity Analysis for Users inMacrocell In this sectionthe capacity of users who are served by the macrocellconsistently before and after range extension is analyzed Forthese users although the serving cell is unchanged the SINRand capacity are impacted anyway by range extension
Suppose the SINR of a user 119906 in this case is representedby 120574119906119872
120574119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(16)
By applying range extension more users are connectedto the picocell and the cell load in the picocell is expectedto increase To facilitate the analysis the interference frompicocell is separated from the total interference in (16)Denoting the picocell as119901 the SINRof user119906 can be rewrittenas
120574119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 = 119901
120588119888119875119888119871119888( 119902119906) + 120588119901119875119901119871119901( 119902119906)
(17)
Denote the SINR of user 119906 after CRE as 1205741015840119906119872
Afterapplying CRE the received signal power from the servingmacrocell can be assumed to be unchanged if we assumethe userrsquos position does not change significantly during thetime of handover executionThe interference from other cellsexcept for the picocell can also be assumed to be unchangedfor simplicity However with CRE the load in the picocell1205881015840
119901 is expected to increase (1205881015840
119901gt 120588119901) Consequently the
interference from picocell will increase eventually The SINRof user 119906 after CRE is then
1205741015840
119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 = 119901
120588119888119875119888119871119888( 119902119906) + 1205881015840
119901119875119901119871119901( 119902119906)
(18)
Therefore the SINR of users served by the macrocell willdecrease due to the increased interference from picocell
1205741015840
119906119872lt 120574119906119872 (19)
Denoting the macrocell as119898 and the capacity of users inmacrocell as119862
119872(119906) the capacity is determined by both SINR
of these users and the load situation in the macrocell
119862119872(119906) = (1 minus 120588
119898) log2(1 + 120574
119906119872) (20)
Denoting the capacity of user 119906 after CRE as 1198621015840119872(119906) we
have
1198621015840
119872(119906) = (1 minus 120588
1015840
119898) log2(1 + 120574
1015840
119906119872) (21)
It is desirable that with CRE the capacity of users stay inmacrocell could increase
1198621015840
119872(119906) gt 119862
119872(119906) (22a)
(1 minus 1205881015840
119898) log2(1 + 120574
1015840
119906119872) gt (1 minus 120588
119898) log2(1 + 120574
119906119872) (22b)
Therefore we can derive the theorem as below to show thecondition that ensures a gain in capacity with CRE for usersstay in macrocell
Theorem2 Thecapacity of users inmacrocell can be increasedif and only if the cell load in themacrocell after CRE satisfies theinequality as follows
1205881015840
119898lt 1 minus (1 minus 120588
119898)
log2(1 + 120574
119906119872)
log2(1 + 120574
1015840
119906119872)
(23)
Theorem 2 shows that only when the cell load of macro-cell decreases low enough to compensate for the loss in SINRthe capacity of these users can increase
More specifically suppose the number of users in themacrocell is 119873
119872 the number of scheduled resource blocks
for each user 119906 is NRB119906 and the total number of available
resource blocks in the cell is NRBavail119898 then the macrocellload 120588
119898is
120588119898=
sum119873119872
119906=1NRB119906
NRBavail119898 (24)
Therefore the capacity of user 119906 is
119862119872(119906) = (1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872) (25)
The number of scheduled resource blocks is usually afunction of SINR Considering a typical traffic type that isnamed as constant bit rate (CBR) service as the example
NRB119906=
119863119906
119877 (120574119906119872) sdot BW (26)
where BW is the bandwidth of one PRB (eg in LTE systemBW is 180 kHz) and 119863
119906is the demanded data rate of user 119906
As shown in (4) 119877(120574) is the achievable data rate that can beestimated according to the Shannon capacity Therefore thecapacity of users stay in macrocell is actually a function ofboth SINRand the number of users inmacrocell Consideringthat the served user number linearly contributes to thecapacity while SINR contributes in logarithmic scale fromthe capacity of macrocell point of view it is desirable tohandover more users out of macrocell However from theoverall capacity point of view it is necessary to consider thecapacity change of both users stay in macrocell and usersstay in CRE region as well as users in picocell which we willanalyze in the following section
6 International Journal of Antennas and Propagation
Suppose the total capacity of users in macrocell is119862119872 we
have
119862119872=
119873119872
sum
119906=1
119862119872(119906)
=
119873119872
sum
119906=1
(1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872)
(27)
223 Capacity Analysis for Users in Picocell The capacity ofusers served by picocell both before and after applying CRE isanalyzed similarly For these users although the serving cellis unchanged the SINR and capacity are eventually impactedby range extension
Suppose the SINR of a user 119906 in this case is presented by120574119906119875
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(28)
To facilitate the analysis the macrocell interference isseparated from the total interference in (28) Denoting themacrocell as119898 the SINR of user 119906 can be rewritten as
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 120588119898119875119898119871119898( 119902119906)
(29)
Denote the SINR of user 119906 after CRE as 1205741015840119906119875
Similarto the analysis for users in macrocell after applying CREthe received signal power from the serving picocell can beassumed to be unchanged and the interference from othercells except for the macrocell can also be assumed to beunchanged for simplicity However with CRE the load in themacrocell denoted by 1205881015840
119898 is changed and usually the load in
macrocell is expected to be reduced (1205881015840119898lt 120588119898) Consequently
the interference from macrocell will decrease The SINR ofpico user u after CRE is then
1205741015840
119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 1205881015840
119898119875119898119871119898( 119902119906)
(30)
Therefore the SINR of users served by picocell increasesdue to the reduced interference as
1205741015840
119906119875gt 120574119906119875 (31)
Denoting the picocell as 119901 and the capacity of user119906 in picocell with and without CRE as 119862
119901(119906) and 1198621015840
119901(119906)
respectively then
119862119875(119906) = (1 minus 120588
119875) log2(1 + 120574
119906119875) (32)
1198621015840
119875(119906) = (1 minus 120588
1015840
119875) log2(1 + 120574
1015840
119906119875) (33)
To achieve capacity gain for users in picocell that is
1198621015840
119875(119906) gt 119862
119875(119906) (34a)
(1 minus 1205881015840
119875) log2(1 + 120574
1015840
119906119875) gt (1 minus 120588
119875) log2(1 + 120574
119906119875) (34b)
Therefore we derive the following theorem which statesthe conditions for users in picocell to achieve capacity gain byCRE
Theorem 3 The capacity of users in picocell can be increasedif and only if the cell load in the picocell after CRE satisfies theinequality as follows
1205881015840
119901lt 1 minus (1 minus 120588
119875)
log2(1 + 120574
119906119875)
log2(1 + 120574
1015840
119906119875)
(35)
FromTheorem 3 it can be concluded that if the increasedcell load in picocell does not exceed the limit as shown ininequality (35) the capacity of users in picocell can increaseOtherwise the capacity of the pico users would decrease
Suppose the number of users in the picocell is 119873119875 the
number of scheduled resource blocks for each user 119906 is NRB119906
and the total number of available resource blocks in thepicocell is NRBavail119901 The picocell load 120588
119901is then
120588119901=
sum119873119875
119906=1NRB119906
NRBavail119901 (36)
And the picocell load after CRE is
1205881015840
119901=
sum119873119875+119873119877
119906=1NRB1015840119906
NRBavail119901 (37)
where 119873119877is the number of users in CRE region and NRB1015840
119906
is the number of resource blocks of user 119906 served by pico Asshown in (26) the number of resource blocks of user119906usuallydepends on its SINR According to the analysis typically theSINR of users in CRE region would decrease while the SINRof users in picocell would increase which results in eitherincreased NRB or decreased NRB However considering theincreased number of served users119873
119877 the cell load in picocell
usually gets difficult to satisfy the inequality of (35) andresults in a decreased capacity
Suppose the total capacity of users in picocell is 119862119875 we
have
119862119875=
119873119875
sum
119906=1
119862119875(119906)
=
119873119875
sum
119906=1
(1 minus
sum119873119875
119906=1NRB119906
NRBavail119875) log2(1 + 120574
119906119875)
(38)
224 Overall Capacity Summarizing capacity of users inmacrocell in picocell and in the CRE region we can get theoverall capacity as
119862total = 119862119872 + 119862119875 + 119862CRE (39)
International Journal of Antennas and Propagation 7
From the analysis above it can be observed that withCRE the overall capacity not only depends on the change ofSINRs of users in macrocell in picocell and in CRE regionbut also depends on the load situation and change in bothmacrocell and picocell CRE may not always give benefit tothe overall capacity in the system Considering an examplethat if the traffic load in macrocell is low and there areenough radio resources that can be allocated to the servedusers the throughput of these users is consequently limited bytheir channel conditions that is SINRs In this case forcingthese users into the picocell by CRE can further damage theSINRs and consequently damage the usersrsquo throughput Onthe other hand if the traffic load in macrocell is high and theusersrsquo throughputs are limited by the available radio resourcehanding over these users to LPN by CRE can improve theusersrsquo throughputs
To get the optimal overall capacity in the next section anadaptive algorithm to optimize theCSO setting via predictionof overall capacity for different CSO values is proposed
3 Capacity Optimization Algorithm withAdaptive Cell Range Control
In this section an adaptive algorithm is proposed to optimizethe overall capacity by adjusting the CSO setting In thisalgorithm the CSO setting is updated periodically At eachupdate instance the current CSO value is updated by a newvalue if the predicted overall capacity can be increased by thisnew attempting valueThe attemptingCSO setting is obtainedby increasing or decreasing a step to the current CSO Thenew CSO value which gives the best overall capacity is thenselected The algorithm is described step by step as in below
31 Step 1 User Grouping with Attempting CSO SettingsSuppose CSO updating step is ΔCSO At each update instance119905 CSO value is updated based on current CSO value CSO
119905minus1
whichwas set at time interval 119905minus1The attemptingCSO valueCSO119905att is obtained by either increasing or decreasingCSO119905minus1
by the updating step ΔCSO as
CSO119905att = CSO119905inc = CSO119905minus1 + ΔCSO (40a)
CSO119905att = CSO119905dec = CSO119905minus1 minus ΔCSO (40b)
Based on the RSRP measurements from both macrocelland picocell for the attempted CSO setting all of the userscan be divided into three groups
Group 1Macro users These users stay in the macrocell bothwith the previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att le RSRP119898
119906
(41)
Group 2 Pico usersThese users stay in the picocell both withthe previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1gt RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(42)
Group 3 Users in CRE region These users will change theirserving cell with new CSO setting Take the case that CSO isincreased as the following example
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(43)
On the basis of the user grouping the capacities of usersin different groups are predicted respectively and the overallcapacity is predicted consequently
32 Step 2 Capacity Prediction with Attempting CSO SettingsIn step 2 the capacities of users in different groups arepredicted respectively according to the predicted SINR andcell load More details can be found below
321 Step 21 Capacity Prediction for Users in CRE RegionTo estimate the new capacity of users in CRE region we needto estimate the new SINR and predict the new cell load inpicocell and in macrocell
Taking the case that the CSO is increased by a step asexample to estimate the new SINR in picocell 120574(119901)
119906CRE we canuse
120574(119901)
119906CRE =RSRP(119901)
(RSRP(119898)120574(119898)119906CRE) + RSRP
(119898)minus RSRP(119901)
(44)
To estimate the new picocell load we need to estimate therequired number of RBs according to the new SINR For CBRservice this can be estimated similarly to (26) by
119873119906=
119863119906
119877 (120574(119901)
119906CRE) sdot BW (45)
For services other than CBR service the new requirednumber of RBs can be estimated according to the relative rela-tionship between the previous SINR in macrocell previousrequired number of RBs inmacrocell and the predicted SINRin picocell
119873119906
new= 119873119906
oldsdot
119877 (120574(119898)
119906CRE)
119877 (120574(119901)
119906CRE) (46)
Based on the prediction of the required number of RBsthe new picocell load can be estimated by assuming therequired number of RBs is unchanged for other users inpicocell
120588119901
new=
sum119873119875
119906=1119873119906+ sum119873119877
119906=1119873119906
new
NRBavail119901
(47)
8 International Journal of Antennas and Propagation
Consequently the new capacity of users newly added toCRE region can be estimated according to the analysis inSection 2
322 Step 22 Capacity Prediction for Macro Users and PicoUsers Formacro users to predict the SINR of users with newCSO the picocell load with new CSO needs to be estimatedThis can be achieved by method described in Step 21
The new cell load in macrocell can be simply estimatedby the new number of served users and assuming that therequired numbers of RBs for macro users are unchanged
The capacity of users in picocell can be estimated simi-larly
33 Optimal CSO Selection and Updating The predictednew capacities of users in different groups are summedtogether to get the overall predicted capacity The predictedoverall capacity is then compared with current capacity If theestimated new capacity is larger than the current capacity theCSO is updated
If 119862 (CSO119905att) gt 119862 (CSO119905minus1) CSO119905 = CSO119905att (48)
Otherwise CSO keeps unchangedIf the CSO is decreased by a step the overall capacity
can be estimated similarly by reversing the calculation shownabove If the new CSO value gives better predicted perfor-mance the CSO value is updated
With the proposed adaptive algorithm the CSO value isadjusted according to the predicted overall capacity in bothmacrocell and picocell By selecting the most suitable CSOvalue the overall capacity is expected to be optimized
4 Numerical Results
The proposed adaptive CSO updating algorithm is evaluatedvia system-level simulations In this section the simulationresults are presented and discussed The simulation assump-tions and parameter settings are illustrated in Table 1
Two typical scenarios are tested in the evaluation Inscenario 1 the users are uniformly distributed in the entiresimulated area LPNs are deployed randomly In scenario 2LPNs are deployed in hotspot areas and the hotspot areastake nearly 60of the usersThe base station deployment anduser distribution in two scenarios are illustrated in Figures 3and 4 To have a clear view only part of the simulated areas isillustrated
The percentages of users belonging to macrocell or LPNsin the two scenarios are illustrated in Figure 5 with differentCSO settings It can be observed that when CSO settingincreases the users belonging to LPNs increase It can beobserved as well that the percentages of users belonging toLPNs in hotspot scenario are much higher than in uniformscenario
Figure 6 shows the change of average cell load withdifferent CSO settings in different scenarios In uniformscenario most of the users are in macrocells and the loadin the macrocell is consequently quite high the cell load inpicocell increases with the increase in the CSO setting In
0
1
2
3
4
5
6
911
15 17
Figure 3 Cell layout and user distribution in uniform scenario
0
1
2
3
4
5
6
911
15 17
Figure 4 Cell layout and user distribution in hotspot scenario
hotspot scenario the average cell load inmacrocells decreaseswith the increase in CSO setting while the average cell load inLPNs increases further with the increased CSO setting
In Figure 7 the average sector throughput gain withdifferent CSO settings is compared with the proposed adap-tive algorithm for uniform scenario The average sectorthroughput with CSO = 0 dB is used as reference to obtainthe average sector throughput gain It can be observed that inthis scenario a higherCSO setting is desirable to offloadmoreusers to LPNs The proposed adaptive algorithm can achievethe similar throughput to the throughput of the optimal CSOsetting
In Figure 8 the average sector throughput gain with dif-ferent CSO settings is compared with the proposed adaptivealgorithm in hotspot scenarioThe average sector throughput
International Journal of Antennas and Propagation 9
0001020304050607080910
Uniform-macroUniform-pico
Hotspot-picoHotspot-macro
CSO 15CSO 10CSO 5CSO 0
Con
nect
ion
ratio
Figure 5 Connection ratio with different CSOs
0 2 4 6 8 10
05
06
07
08
09
10
Uniform-macroUniform-pico
Hotspot-macroHotspot-pico
CSO (dB)
Nor
mal
ized
cell
load
Figure 6 Cell load with different CSOs
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 7 Average sector throughput gain in uniform scenario
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 8 Average sector throughput gain in hotspot scenario
with CSO = 0 dB is used as reference to obtain the averagesector throughput gain In this scenario CSO of 10 dB isdesirable to balance the offloading gain and SINR loss Theproposed algorithm achieves the similar performance to theoptimal CSO setting
In different scenarios the optimal CSO setting is differ-ent However the proposed adaptive algorithm can achievesimilar performance with the optimal CSO setting in bothscenarios
5 Conclusion
As an attractivemeans of expandingmobile network capacityheterogeneous network has been included in LTE-advancedCell range extension is an important feature in heterogeneousnetwork to improve the utilization of the resource of lowpower nodes However the users in the cell range extensionarea usually suffer from high interference from macrocellsTo avoid the unnecessary interference it is desirable tocarefully consider the optimal cell selection offset value Inthis paper the capacity of users in macrocell and small cellsis analyzed thoroughly The conditions to improve capacityvia cell range extension are derived respectively for usersboth in macrocell and in small cell Based on the analysis anadaptive small cell coverage control algorithm is proposedWith the proposed adaptive algorithm the cell selectionoffset is updated periodically based on the prediction of theoverall system capacityThe algorithm is evaluated by system-level simulations and the results indicate that with theproposed algorithm a nearly optimal system performancecan be achieved in all tested load cases
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
10 International Journal of Antennas and Propagation
Acknowledgments
This work was supported by the State Major Science andTechnology Special Projects (Grant no 2013ZX03001026-001) and the Fundamental Research Funds for the CentralUniversities (Grant no 2014RC0107)
References
[1] A Damnjanovic J Montojo Y Wei et al ldquoA survey on 3GPPheterogeneous networksrdquo IEEE Wireless Communications vol18 no 3 pp 10ndash21 2011
[2] E Dahlman S Parkvall and J Skold 4G LTELTE-Advanced forMobile Broadband Academic Press 2011
[3] A Ghosh R Ratasuk B Mondal N Mangalvedhe and TThomas ldquoLTE-advanced next-generation wireless broadbandtechnologyrdquo IEEE Wireless Communications vol 17 no 3 pp10ndash22 2010
[4] X Zhang XGuW Li L Zhang J Shen andYWan ldquoThe studyof indoor and field trials on 2times8MIMOarchitecture in TD-LTEnetworkrdquo International Journal of Antennas and Propagationsvol 2013 Article ID 181579 9 pages 2013
[5] G Yuan X Zhang WWang and Y Yang ldquoCarrier aggregationfor LTE-advancedmobile communication systemsrdquo IEEE Com-munications Magazine vol 48 no 2 pp 88ndash93 2010
[6] R Irmer H Droste P Marsch et al ldquoCoordinated multipointconcepts performance and field trial resultsrdquo IEEE Communi-cations Magazine vol 49 no 2 pp 102ndash111 2011
[7] I Siomina and Y Di ldquoLoad balancing in heterogeneous LTErange optimization via cell offset and load-coupling character-izationrdquo in Proceedings of the IEEE International Conference onCommunications pp 1357ndash1361 2012
[8] The 3rd Generation Partnership Project (3GPP) ldquoFeasibilitystudy for further advancements for E-UTRA (LTEAdvanced)(Release 10)rdquo Tech Rep TR 36912 2011
[9] The 3rd Generation Partnership Project(3GPP) ldquoSystem per-formance of heterogeneous networks with range expansionrdquoTech Rep R1-101203 Samsung 2010
[10] I Guvenc ldquoCapacity and fairness analysis of heterogeneous net-works with range expansion and interference coordinationrdquoIEEECommunications Letters vol 15 no 10 pp 1084ndash1087 2011
[11] H-S Jo Y J Sang P Xia and J G Andrews ldquoHeterogeneouscellular networks with flexible cell association a comprehensivedownlink SINR analysisrdquo IEEE Transactions on Wireless Com-munications vol 11 no 10 pp 3484ndash3495 2012
[12] ldquoSelf-optimisationand self-configuration in wireless networksrdquoSOCRATES European Research Project httpwwwfp7-socratesorg
[13] The 3rd Generation Partnership Project(3GPP) ldquoSelf-config-uring and self-optimizing network use cases and solutionsrdquoTech Rep TR 36902 2009
[14] ldquoUse cases related to Self Organizing network Overall descrip-tionrdquoNext generationMobileNetworks httpwwwngmnorg
[15] X Chu and D Lopez-Perez Heterogeneous Cellular NetworksTheory Simulation and Deployment Cambridge UniversityPress 2013
[16] P Tian H Tian J Zhu L Chen and X She ldquoAn adaptive biasconfiguration strategy for range extension in LTE-Advancedheterogeneous networksrdquo inProceedings of the IET InternationalConference on Communication Technology and Application(ICCTA rsquo11) pp 336ndash340 2011
[17] K Kikuchi and H Otsuka ldquoProposal of adaptive control CREin heterogeneous networksrdquo in Proceedings of the IEEE Inter-national Symposium on Personal Indoor and Mobile RadioCommunications (PIMRC rsquo12) pp 910ndash914 2012
[18] I Viering M Dottling and A Lobinger ldquoA mathematical per-spective of self-optimizing wireless networksrdquo in Proceedings ofthe IEEE International Conference on Communications (ICCrsquo09) pp 1ndash6 June 2009
[19] A Lobinger S Stefanski T Jansen and I Balan ldquoLoad balanc-ing in downlink LTE self-optimizing networksrdquo in Proceedingsof the IEEE 71st Vehicular Technology Conference (VTC rsquo10) pp1ndash5 May 2010
[20] ldquoGuidelines for evaluation of radio transmission technologiesfor IMT-2000rdquo Recommendation ITU-R M1225 1997
[21] The 3rd Generation Partnership Project (3GPP) ldquoFurtheradvancements for E-UTRAN physical layer aspects (Release9)rdquo Tech Rep TS 36814 2010
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Antennas and Propagation 3
cell selection usually the UE measures the reference signalreceived power (RSRP) from each cell and selects the servingcell with the maximum received signal power as shown by
119878 (119906) = argmax119888
(RSRP119906119888) = argmax
119888
(119875119888sdot 119871119888( 119902119906)) (2)
In (2) UE119906 selects cell c as its serving cell 119878(119906) because theRSRP from cell 119888 is the maximum among the measurementsfrom different cells
In homogeneous network the cell with the maximumRSRP is usually the cell with the best path-gain to thisUE however in heterogeneous network where the deployedbase stations have different transmit powers the cell withmaximum RSRP may no longer be the cell with the bestpath-gain to the UE As shown in (2) the RSRP of user 119906received from cell 119888 is the transmit power of cell 119888multipliedby path-gain (119871
119888) between cell 119888 and user 119906 (at the position
119902119906) Considering a typical transmit power of macrocell and
picocell as 43 dBm and 30 dBm respectively only UE whosepath-gain to the picocell is 13 dB better than the path-gain tomacrocell can select the picocell as its serving cell This leadsto a small number of served users in LPNs
To cope with this issue CRE is proposed for heteroge-neous network in 3GPPWithCRE a cell specific bias CSO isadded to the measured RSRP during cell selection procedureas shown in
119878 (119906) = argmax119888
(RSRP119906119888+ CSO119888) (3)
TheCSO value is usually zero formacrocell and a positivevalue for a LPN
22 Capacity Analysis with Cell Range Extension Assumingideal link adaptation which means that the most suitablemodulation and coding scheme (MCS) is selected accordingto the SINR of user 119906 the achievable data rate can berepresented by the Shannon capacity as
119877 (120574119906) = log
2(1 + 120574
119906) (4)
Here 119877(120574119906) is the achievable data rate if user 119906 is sched-
uled Considering the normal case that more than one user isconnecting to the same cell and sharing the resource in thesame cell and by assuming a round robin scheduling strategythe achievable capacity can be represented by
119862 (119906) = (1 minus 120588119878(119906)) log2(1 + 120574
119906) (5)
where120588119878(119906)
is the load in the serving cell 119878(119906) Intuitively usersin a cell with high load (large 120588
119888) have lower probability to get
chance to be scheduled and even if the SINR of this user isgood the final capacity is still not high
221 Capacity Analysis for Users in CRE Region In this sec-tion capacity of users in the CRE region is analyzed For theseusers before a CSO value is applied the serving cell of theseusers is the macrocell and after the CSO value is applied theserving cell of these users is the LPNWe analyze the capacitychange of these users with and without CRE
Denote the capacity of user 119906 in CRE region by 119862CRE(119906)For these users the serving cell 119878(119906) will change with andwithout CRE Represent the serving cell before applying CSOas 119878(119898)(119906) and the serving cell after applying CSO as 119878(119901)(119906)The SINR of user 119906 before applying CSO is 120574(119898)
119906CRE
120574(119898)
119906CRE =119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906)
119873 + sum119888 = 119878(119898)(119906)120588119888119875119888119871119888( 119902119906)
(6)
To facilitate further analysis the interference from thepicocell which is the serving cell after applying CSO isseparated from the overall interference Therefore the SINRof user 119906 before applying CSO is
120574(119898)
119906CRE =119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906)
119873 + sum119888 = 119878(119898)(119906)120588119888119875119888119871119888( 119902119906)
= (119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906))
times(119873 + sum
119888 = 119878(119898)(119906)
119888 = 119878(119901)(119906)
120588119888119875119888119871119888( 119902119906)
+ 120588119878(119901)(119906)119875119878(119901)(119906)119871119878(119901)(119906)( 119902119906))
minus1
(7)
Suppose the SINR of user 119906 after applying CSO is 120574(119901)119906CRE
120574(119901)
119906CRE =119875119883(119901)(119906)119871119883(119901)(119906)( 119902119906)
119873 + sum119888 =119883(119901)(119906)120588119888119875119888119871119888( 119902119906)
(8)
With CRE the received power from the previous servingcell becomes the interference Similarly the interference fromthe macrocell which is the serving cell before applying CSOis extracted from the overall interference for analysis TheSINR of user u after applying CSO could be written as
120574(119901)
119906CRE =119875119878(119901)(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
= (119875119878(119901)(119906)119871119878(119901)(119906)( 119902119906))
times(119873 + sum
119888 = 119878(119898)(119906)
119888 = 119878(119901)(119906)
120588119888119875119888119871119888( 119902119906)
+120588119878(119898)(119906)119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906))
minus1
(9)
4 International Journal of Antennas and Propagation
Table 1 System evaluation assumptions and parameters
Parameters Valuesdescriptions
Macro base stationdeployment
19 3-sector cells with wrap-aroundSite to site distance 500mHeight 40m above groundTx power 46 dBm
LPN deployment
6 LPNs per macrocell on averageHeight 5m above groundTx power 30 dBmCell radius 30m
Propagation model ITU UMa [20]
Antenna model Macro 3GPP [21]LPN dipole
Transmission scheme SU-MIMO 2Tx 2RxScheduler Proportional fairTraffic type DownloadNumber of users permacrocell 75
User distribution Uniform or in hotspot
For a user in the cell range extension area although theselected serving cell is the picocell the received power frommacro base station is still larger than received signal powerfrom picocell This means that for these users it is true that
119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906) ge 119875119878(119901)(119906)119871119878(119901)(119906)( 119902119906) (10)
Considering the most common case that the load in themacrocell is not lower than that in the picocell comparing(7) and (9) it can be concluded that the UE who are ldquoforcedrdquointo picocells by range extension will suffer from lower SINRThat is
120574(119901)
119906CRE le 120574(119898)
119906CRE (11)
For example Figure 2 clearly shows the SINR loss of usersin the cell range extension area In Figure 2 the cell selectionoffset is 6 dB and more simulation settings can be found inTable 1
Based on the analysis of the change of SINRof users in cellrange extension region the capacity of these users with andwithout cell range extension is further analyzed Suppose thecapacity of user 119906 before and after applying CSO is 119862(119898)CRE(119906)
and 119862(119901)CRE(119906) respectively
119862(119898)
CRE (119906) = (1 minus 120588119878(119898)(119906)) log2 (1 + 120574(119898)
119906CRE) (12a)
119862(119901)
CRE (119906) = (1 minus 120588119878(119901)(119906)) log2 (1 + 120574(119901)
119906CRE) (12b)
Comparing (12a) and (12b) it can be seen that althoughthe SINR is expected to be reduced (120574(119901)
119906CRE le 120574(119898)
119906CRE) forusers in CRE region after range extension the capacity ofthese users depends on the cell load of their serving cells aswell If the cell load in the picocell is below the cell load in
0 5 10 15 20 250
10
20
30
40
50
60
70
80
90
100
CDF
()
SINR of users in CRE region
SINR (dB)
wo CREw CRE
minus15 minus10 minus5
Figure 2 SINR of users in range extension area
the macrocell by certain extent although there is an expectedSINR loss the capacity of users in the CRE region can stillexpect an increase
119862(119901)
CRE (119906) gt 119862(119898)
CRE (119906) (13a)
(1minus120588119878(119901)(119906)) log2(1+120574(119901)
119906CRE) gt (1minus120588119878(119898)(119906)) log2 (1+120574(119898)
119906CRE)
(13b)
Therefore we can derive the theorem below to show thecondition that ensures a gain in capacity of users in CREregion
Theorem 1 The capacity of users in region of CRE can beincreased if and only if the cell load in the original servingmacrocell before CRE satisfies the inequality as below
120588119878(119898)(119906)gt 1 minus (1 minus 120588
119878(119901)(119906))
log2(1 + 120574
(119901)
119906CRE)
log2(1 + 120574
(119898)
119906CRE) (14)
Comparing the role of SINR and cell load in capacityit can be seen that the cell load linearly contributes tothe capacity while SINR contributes in logarithmic scaleTherefore the contribution from cell load plays an evenmoreimportant role in capacity than SINR does and needs to betreated seriously
To summarize on one hand for users in the CREregion usually the SINR decreases due to the suboptimalreceived signal power and high interference from macrocellon the other hand CRE may also bring benefit in terms ofavailable resource (eg frequency time and power) to theseusers More resource to a user means for example moreopportunity to be scheduled by the networkwhich eventuallycontributes to the user capacity
International Journal of Antennas and Propagation 5
Suppose the total capacity of users in CRE region is119862CREand the number of users in the CRE region is119873
119877 we have
119862CRE =
119873119877
sum
119906=1
119862CRE (119906) (15)
222 Capacity Analysis for Users inMacrocell In this sectionthe capacity of users who are served by the macrocellconsistently before and after range extension is analyzed Forthese users although the serving cell is unchanged the SINRand capacity are impacted anyway by range extension
Suppose the SINR of a user 119906 in this case is representedby 120574119906119872
120574119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(16)
By applying range extension more users are connectedto the picocell and the cell load in the picocell is expectedto increase To facilitate the analysis the interference frompicocell is separated from the total interference in (16)Denoting the picocell as119901 the SINRof user119906 can be rewrittenas
120574119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 = 119901
120588119888119875119888119871119888( 119902119906) + 120588119901119875119901119871119901( 119902119906)
(17)
Denote the SINR of user 119906 after CRE as 1205741015840119906119872
Afterapplying CRE the received signal power from the servingmacrocell can be assumed to be unchanged if we assumethe userrsquos position does not change significantly during thetime of handover executionThe interference from other cellsexcept for the picocell can also be assumed to be unchangedfor simplicity However with CRE the load in the picocell1205881015840
119901 is expected to increase (1205881015840
119901gt 120588119901) Consequently the
interference from picocell will increase eventually The SINRof user 119906 after CRE is then
1205741015840
119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 = 119901
120588119888119875119888119871119888( 119902119906) + 1205881015840
119901119875119901119871119901( 119902119906)
(18)
Therefore the SINR of users served by the macrocell willdecrease due to the increased interference from picocell
1205741015840
119906119872lt 120574119906119872 (19)
Denoting the macrocell as119898 and the capacity of users inmacrocell as119862
119872(119906) the capacity is determined by both SINR
of these users and the load situation in the macrocell
119862119872(119906) = (1 minus 120588
119898) log2(1 + 120574
119906119872) (20)
Denoting the capacity of user 119906 after CRE as 1198621015840119872(119906) we
have
1198621015840
119872(119906) = (1 minus 120588
1015840
119898) log2(1 + 120574
1015840
119906119872) (21)
It is desirable that with CRE the capacity of users stay inmacrocell could increase
1198621015840
119872(119906) gt 119862
119872(119906) (22a)
(1 minus 1205881015840
119898) log2(1 + 120574
1015840
119906119872) gt (1 minus 120588
119898) log2(1 + 120574
119906119872) (22b)
Therefore we can derive the theorem as below to show thecondition that ensures a gain in capacity with CRE for usersstay in macrocell
Theorem2 Thecapacity of users inmacrocell can be increasedif and only if the cell load in themacrocell after CRE satisfies theinequality as follows
1205881015840
119898lt 1 minus (1 minus 120588
119898)
log2(1 + 120574
119906119872)
log2(1 + 120574
1015840
119906119872)
(23)
Theorem 2 shows that only when the cell load of macro-cell decreases low enough to compensate for the loss in SINRthe capacity of these users can increase
More specifically suppose the number of users in themacrocell is 119873
119872 the number of scheduled resource blocks
for each user 119906 is NRB119906 and the total number of available
resource blocks in the cell is NRBavail119898 then the macrocellload 120588
119898is
120588119898=
sum119873119872
119906=1NRB119906
NRBavail119898 (24)
Therefore the capacity of user 119906 is
119862119872(119906) = (1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872) (25)
The number of scheduled resource blocks is usually afunction of SINR Considering a typical traffic type that isnamed as constant bit rate (CBR) service as the example
NRB119906=
119863119906
119877 (120574119906119872) sdot BW (26)
where BW is the bandwidth of one PRB (eg in LTE systemBW is 180 kHz) and 119863
119906is the demanded data rate of user 119906
As shown in (4) 119877(120574) is the achievable data rate that can beestimated according to the Shannon capacity Therefore thecapacity of users stay in macrocell is actually a function ofboth SINRand the number of users inmacrocell Consideringthat the served user number linearly contributes to thecapacity while SINR contributes in logarithmic scale fromthe capacity of macrocell point of view it is desirable tohandover more users out of macrocell However from theoverall capacity point of view it is necessary to consider thecapacity change of both users stay in macrocell and usersstay in CRE region as well as users in picocell which we willanalyze in the following section
6 International Journal of Antennas and Propagation
Suppose the total capacity of users in macrocell is119862119872 we
have
119862119872=
119873119872
sum
119906=1
119862119872(119906)
=
119873119872
sum
119906=1
(1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872)
(27)
223 Capacity Analysis for Users in Picocell The capacity ofusers served by picocell both before and after applying CRE isanalyzed similarly For these users although the serving cellis unchanged the SINR and capacity are eventually impactedby range extension
Suppose the SINR of a user 119906 in this case is presented by120574119906119875
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(28)
To facilitate the analysis the macrocell interference isseparated from the total interference in (28) Denoting themacrocell as119898 the SINR of user 119906 can be rewritten as
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 120588119898119875119898119871119898( 119902119906)
(29)
Denote the SINR of user 119906 after CRE as 1205741015840119906119875
Similarto the analysis for users in macrocell after applying CREthe received signal power from the serving picocell can beassumed to be unchanged and the interference from othercells except for the macrocell can also be assumed to beunchanged for simplicity However with CRE the load in themacrocell denoted by 1205881015840
119898 is changed and usually the load in
macrocell is expected to be reduced (1205881015840119898lt 120588119898) Consequently
the interference from macrocell will decrease The SINR ofpico user u after CRE is then
1205741015840
119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 1205881015840
119898119875119898119871119898( 119902119906)
(30)
Therefore the SINR of users served by picocell increasesdue to the reduced interference as
1205741015840
119906119875gt 120574119906119875 (31)
Denoting the picocell as 119901 and the capacity of user119906 in picocell with and without CRE as 119862
119901(119906) and 1198621015840
119901(119906)
respectively then
119862119875(119906) = (1 minus 120588
119875) log2(1 + 120574
119906119875) (32)
1198621015840
119875(119906) = (1 minus 120588
1015840
119875) log2(1 + 120574
1015840
119906119875) (33)
To achieve capacity gain for users in picocell that is
1198621015840
119875(119906) gt 119862
119875(119906) (34a)
(1 minus 1205881015840
119875) log2(1 + 120574
1015840
119906119875) gt (1 minus 120588
119875) log2(1 + 120574
119906119875) (34b)
Therefore we derive the following theorem which statesthe conditions for users in picocell to achieve capacity gain byCRE
Theorem 3 The capacity of users in picocell can be increasedif and only if the cell load in the picocell after CRE satisfies theinequality as follows
1205881015840
119901lt 1 minus (1 minus 120588
119875)
log2(1 + 120574
119906119875)
log2(1 + 120574
1015840
119906119875)
(35)
FromTheorem 3 it can be concluded that if the increasedcell load in picocell does not exceed the limit as shown ininequality (35) the capacity of users in picocell can increaseOtherwise the capacity of the pico users would decrease
Suppose the number of users in the picocell is 119873119875 the
number of scheduled resource blocks for each user 119906 is NRB119906
and the total number of available resource blocks in thepicocell is NRBavail119901 The picocell load 120588
119901is then
120588119901=
sum119873119875
119906=1NRB119906
NRBavail119901 (36)
And the picocell load after CRE is
1205881015840
119901=
sum119873119875+119873119877
119906=1NRB1015840119906
NRBavail119901 (37)
where 119873119877is the number of users in CRE region and NRB1015840
119906
is the number of resource blocks of user 119906 served by pico Asshown in (26) the number of resource blocks of user119906usuallydepends on its SINR According to the analysis typically theSINR of users in CRE region would decrease while the SINRof users in picocell would increase which results in eitherincreased NRB or decreased NRB However considering theincreased number of served users119873
119877 the cell load in picocell
usually gets difficult to satisfy the inequality of (35) andresults in a decreased capacity
Suppose the total capacity of users in picocell is 119862119875 we
have
119862119875=
119873119875
sum
119906=1
119862119875(119906)
=
119873119875
sum
119906=1
(1 minus
sum119873119875
119906=1NRB119906
NRBavail119875) log2(1 + 120574
119906119875)
(38)
224 Overall Capacity Summarizing capacity of users inmacrocell in picocell and in the CRE region we can get theoverall capacity as
119862total = 119862119872 + 119862119875 + 119862CRE (39)
International Journal of Antennas and Propagation 7
From the analysis above it can be observed that withCRE the overall capacity not only depends on the change ofSINRs of users in macrocell in picocell and in CRE regionbut also depends on the load situation and change in bothmacrocell and picocell CRE may not always give benefit tothe overall capacity in the system Considering an examplethat if the traffic load in macrocell is low and there areenough radio resources that can be allocated to the servedusers the throughput of these users is consequently limited bytheir channel conditions that is SINRs In this case forcingthese users into the picocell by CRE can further damage theSINRs and consequently damage the usersrsquo throughput Onthe other hand if the traffic load in macrocell is high and theusersrsquo throughputs are limited by the available radio resourcehanding over these users to LPN by CRE can improve theusersrsquo throughputs
To get the optimal overall capacity in the next section anadaptive algorithm to optimize theCSO setting via predictionof overall capacity for different CSO values is proposed
3 Capacity Optimization Algorithm withAdaptive Cell Range Control
In this section an adaptive algorithm is proposed to optimizethe overall capacity by adjusting the CSO setting In thisalgorithm the CSO setting is updated periodically At eachupdate instance the current CSO value is updated by a newvalue if the predicted overall capacity can be increased by thisnew attempting valueThe attemptingCSO setting is obtainedby increasing or decreasing a step to the current CSO Thenew CSO value which gives the best overall capacity is thenselected The algorithm is described step by step as in below
31 Step 1 User Grouping with Attempting CSO SettingsSuppose CSO updating step is ΔCSO At each update instance119905 CSO value is updated based on current CSO value CSO
119905minus1
whichwas set at time interval 119905minus1The attemptingCSO valueCSO119905att is obtained by either increasing or decreasingCSO119905minus1
by the updating step ΔCSO as
CSO119905att = CSO119905inc = CSO119905minus1 + ΔCSO (40a)
CSO119905att = CSO119905dec = CSO119905minus1 minus ΔCSO (40b)
Based on the RSRP measurements from both macrocelland picocell for the attempted CSO setting all of the userscan be divided into three groups
Group 1Macro users These users stay in the macrocell bothwith the previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att le RSRP119898
119906
(41)
Group 2 Pico usersThese users stay in the picocell both withthe previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1gt RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(42)
Group 3 Users in CRE region These users will change theirserving cell with new CSO setting Take the case that CSO isincreased as the following example
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(43)
On the basis of the user grouping the capacities of usersin different groups are predicted respectively and the overallcapacity is predicted consequently
32 Step 2 Capacity Prediction with Attempting CSO SettingsIn step 2 the capacities of users in different groups arepredicted respectively according to the predicted SINR andcell load More details can be found below
321 Step 21 Capacity Prediction for Users in CRE RegionTo estimate the new capacity of users in CRE region we needto estimate the new SINR and predict the new cell load inpicocell and in macrocell
Taking the case that the CSO is increased by a step asexample to estimate the new SINR in picocell 120574(119901)
119906CRE we canuse
120574(119901)
119906CRE =RSRP(119901)
(RSRP(119898)120574(119898)119906CRE) + RSRP
(119898)minus RSRP(119901)
(44)
To estimate the new picocell load we need to estimate therequired number of RBs according to the new SINR For CBRservice this can be estimated similarly to (26) by
119873119906=
119863119906
119877 (120574(119901)
119906CRE) sdot BW (45)
For services other than CBR service the new requirednumber of RBs can be estimated according to the relative rela-tionship between the previous SINR in macrocell previousrequired number of RBs inmacrocell and the predicted SINRin picocell
119873119906
new= 119873119906
oldsdot
119877 (120574(119898)
119906CRE)
119877 (120574(119901)
119906CRE) (46)
Based on the prediction of the required number of RBsthe new picocell load can be estimated by assuming therequired number of RBs is unchanged for other users inpicocell
120588119901
new=
sum119873119875
119906=1119873119906+ sum119873119877
119906=1119873119906
new
NRBavail119901
(47)
8 International Journal of Antennas and Propagation
Consequently the new capacity of users newly added toCRE region can be estimated according to the analysis inSection 2
322 Step 22 Capacity Prediction for Macro Users and PicoUsers Formacro users to predict the SINR of users with newCSO the picocell load with new CSO needs to be estimatedThis can be achieved by method described in Step 21
The new cell load in macrocell can be simply estimatedby the new number of served users and assuming that therequired numbers of RBs for macro users are unchanged
The capacity of users in picocell can be estimated simi-larly
33 Optimal CSO Selection and Updating The predictednew capacities of users in different groups are summedtogether to get the overall predicted capacity The predictedoverall capacity is then compared with current capacity If theestimated new capacity is larger than the current capacity theCSO is updated
If 119862 (CSO119905att) gt 119862 (CSO119905minus1) CSO119905 = CSO119905att (48)
Otherwise CSO keeps unchangedIf the CSO is decreased by a step the overall capacity
can be estimated similarly by reversing the calculation shownabove If the new CSO value gives better predicted perfor-mance the CSO value is updated
With the proposed adaptive algorithm the CSO value isadjusted according to the predicted overall capacity in bothmacrocell and picocell By selecting the most suitable CSOvalue the overall capacity is expected to be optimized
4 Numerical Results
The proposed adaptive CSO updating algorithm is evaluatedvia system-level simulations In this section the simulationresults are presented and discussed The simulation assump-tions and parameter settings are illustrated in Table 1
Two typical scenarios are tested in the evaluation Inscenario 1 the users are uniformly distributed in the entiresimulated area LPNs are deployed randomly In scenario 2LPNs are deployed in hotspot areas and the hotspot areastake nearly 60of the usersThe base station deployment anduser distribution in two scenarios are illustrated in Figures 3and 4 To have a clear view only part of the simulated areas isillustrated
The percentages of users belonging to macrocell or LPNsin the two scenarios are illustrated in Figure 5 with differentCSO settings It can be observed that when CSO settingincreases the users belonging to LPNs increase It can beobserved as well that the percentages of users belonging toLPNs in hotspot scenario are much higher than in uniformscenario
Figure 6 shows the change of average cell load withdifferent CSO settings in different scenarios In uniformscenario most of the users are in macrocells and the loadin the macrocell is consequently quite high the cell load inpicocell increases with the increase in the CSO setting In
0
1
2
3
4
5
6
911
15 17
Figure 3 Cell layout and user distribution in uniform scenario
0
1
2
3
4
5
6
911
15 17
Figure 4 Cell layout and user distribution in hotspot scenario
hotspot scenario the average cell load inmacrocells decreaseswith the increase in CSO setting while the average cell load inLPNs increases further with the increased CSO setting
In Figure 7 the average sector throughput gain withdifferent CSO settings is compared with the proposed adap-tive algorithm for uniform scenario The average sectorthroughput with CSO = 0 dB is used as reference to obtainthe average sector throughput gain It can be observed that inthis scenario a higherCSO setting is desirable to offloadmoreusers to LPNs The proposed adaptive algorithm can achievethe similar throughput to the throughput of the optimal CSOsetting
In Figure 8 the average sector throughput gain with dif-ferent CSO settings is compared with the proposed adaptivealgorithm in hotspot scenarioThe average sector throughput
International Journal of Antennas and Propagation 9
0001020304050607080910
Uniform-macroUniform-pico
Hotspot-picoHotspot-macro
CSO 15CSO 10CSO 5CSO 0
Con
nect
ion
ratio
Figure 5 Connection ratio with different CSOs
0 2 4 6 8 10
05
06
07
08
09
10
Uniform-macroUniform-pico
Hotspot-macroHotspot-pico
CSO (dB)
Nor
mal
ized
cell
load
Figure 6 Cell load with different CSOs
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 7 Average sector throughput gain in uniform scenario
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 8 Average sector throughput gain in hotspot scenario
with CSO = 0 dB is used as reference to obtain the averagesector throughput gain In this scenario CSO of 10 dB isdesirable to balance the offloading gain and SINR loss Theproposed algorithm achieves the similar performance to theoptimal CSO setting
In different scenarios the optimal CSO setting is differ-ent However the proposed adaptive algorithm can achievesimilar performance with the optimal CSO setting in bothscenarios
5 Conclusion
As an attractivemeans of expandingmobile network capacityheterogeneous network has been included in LTE-advancedCell range extension is an important feature in heterogeneousnetwork to improve the utilization of the resource of lowpower nodes However the users in the cell range extensionarea usually suffer from high interference from macrocellsTo avoid the unnecessary interference it is desirable tocarefully consider the optimal cell selection offset value Inthis paper the capacity of users in macrocell and small cellsis analyzed thoroughly The conditions to improve capacityvia cell range extension are derived respectively for usersboth in macrocell and in small cell Based on the analysis anadaptive small cell coverage control algorithm is proposedWith the proposed adaptive algorithm the cell selectionoffset is updated periodically based on the prediction of theoverall system capacityThe algorithm is evaluated by system-level simulations and the results indicate that with theproposed algorithm a nearly optimal system performancecan be achieved in all tested load cases
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
10 International Journal of Antennas and Propagation
Acknowledgments
This work was supported by the State Major Science andTechnology Special Projects (Grant no 2013ZX03001026-001) and the Fundamental Research Funds for the CentralUniversities (Grant no 2014RC0107)
References
[1] A Damnjanovic J Montojo Y Wei et al ldquoA survey on 3GPPheterogeneous networksrdquo IEEE Wireless Communications vol18 no 3 pp 10ndash21 2011
[2] E Dahlman S Parkvall and J Skold 4G LTELTE-Advanced forMobile Broadband Academic Press 2011
[3] A Ghosh R Ratasuk B Mondal N Mangalvedhe and TThomas ldquoLTE-advanced next-generation wireless broadbandtechnologyrdquo IEEE Wireless Communications vol 17 no 3 pp10ndash22 2010
[4] X Zhang XGuW Li L Zhang J Shen andYWan ldquoThe studyof indoor and field trials on 2times8MIMOarchitecture in TD-LTEnetworkrdquo International Journal of Antennas and Propagationsvol 2013 Article ID 181579 9 pages 2013
[5] G Yuan X Zhang WWang and Y Yang ldquoCarrier aggregationfor LTE-advancedmobile communication systemsrdquo IEEE Com-munications Magazine vol 48 no 2 pp 88ndash93 2010
[6] R Irmer H Droste P Marsch et al ldquoCoordinated multipointconcepts performance and field trial resultsrdquo IEEE Communi-cations Magazine vol 49 no 2 pp 102ndash111 2011
[7] I Siomina and Y Di ldquoLoad balancing in heterogeneous LTErange optimization via cell offset and load-coupling character-izationrdquo in Proceedings of the IEEE International Conference onCommunications pp 1357ndash1361 2012
[8] The 3rd Generation Partnership Project (3GPP) ldquoFeasibilitystudy for further advancements for E-UTRA (LTEAdvanced)(Release 10)rdquo Tech Rep TR 36912 2011
[9] The 3rd Generation Partnership Project(3GPP) ldquoSystem per-formance of heterogeneous networks with range expansionrdquoTech Rep R1-101203 Samsung 2010
[10] I Guvenc ldquoCapacity and fairness analysis of heterogeneous net-works with range expansion and interference coordinationrdquoIEEECommunications Letters vol 15 no 10 pp 1084ndash1087 2011
[11] H-S Jo Y J Sang P Xia and J G Andrews ldquoHeterogeneouscellular networks with flexible cell association a comprehensivedownlink SINR analysisrdquo IEEE Transactions on Wireless Com-munications vol 11 no 10 pp 3484ndash3495 2012
[12] ldquoSelf-optimisationand self-configuration in wireless networksrdquoSOCRATES European Research Project httpwwwfp7-socratesorg
[13] The 3rd Generation Partnership Project(3GPP) ldquoSelf-config-uring and self-optimizing network use cases and solutionsrdquoTech Rep TR 36902 2009
[14] ldquoUse cases related to Self Organizing network Overall descrip-tionrdquoNext generationMobileNetworks httpwwwngmnorg
[15] X Chu and D Lopez-Perez Heterogeneous Cellular NetworksTheory Simulation and Deployment Cambridge UniversityPress 2013
[16] P Tian H Tian J Zhu L Chen and X She ldquoAn adaptive biasconfiguration strategy for range extension in LTE-Advancedheterogeneous networksrdquo inProceedings of the IET InternationalConference on Communication Technology and Application(ICCTA rsquo11) pp 336ndash340 2011
[17] K Kikuchi and H Otsuka ldquoProposal of adaptive control CREin heterogeneous networksrdquo in Proceedings of the IEEE Inter-national Symposium on Personal Indoor and Mobile RadioCommunications (PIMRC rsquo12) pp 910ndash914 2012
[18] I Viering M Dottling and A Lobinger ldquoA mathematical per-spective of self-optimizing wireless networksrdquo in Proceedings ofthe IEEE International Conference on Communications (ICCrsquo09) pp 1ndash6 June 2009
[19] A Lobinger S Stefanski T Jansen and I Balan ldquoLoad balanc-ing in downlink LTE self-optimizing networksrdquo in Proceedingsof the IEEE 71st Vehicular Technology Conference (VTC rsquo10) pp1ndash5 May 2010
[20] ldquoGuidelines for evaluation of radio transmission technologiesfor IMT-2000rdquo Recommendation ITU-R M1225 1997
[21] The 3rd Generation Partnership Project (3GPP) ldquoFurtheradvancements for E-UTRAN physical layer aspects (Release9)rdquo Tech Rep TS 36814 2010
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
4 International Journal of Antennas and Propagation
Table 1 System evaluation assumptions and parameters
Parameters Valuesdescriptions
Macro base stationdeployment
19 3-sector cells with wrap-aroundSite to site distance 500mHeight 40m above groundTx power 46 dBm
LPN deployment
6 LPNs per macrocell on averageHeight 5m above groundTx power 30 dBmCell radius 30m
Propagation model ITU UMa [20]
Antenna model Macro 3GPP [21]LPN dipole
Transmission scheme SU-MIMO 2Tx 2RxScheduler Proportional fairTraffic type DownloadNumber of users permacrocell 75
User distribution Uniform or in hotspot
For a user in the cell range extension area although theselected serving cell is the picocell the received power frommacro base station is still larger than received signal powerfrom picocell This means that for these users it is true that
119875119878(119898)(119906)119871119878(119898)(119906)( 119902119906) ge 119875119878(119901)(119906)119871119878(119901)(119906)( 119902119906) (10)
Considering the most common case that the load in themacrocell is not lower than that in the picocell comparing(7) and (9) it can be concluded that the UE who are ldquoforcedrdquointo picocells by range extension will suffer from lower SINRThat is
120574(119901)
119906CRE le 120574(119898)
119906CRE (11)
For example Figure 2 clearly shows the SINR loss of usersin the cell range extension area In Figure 2 the cell selectionoffset is 6 dB and more simulation settings can be found inTable 1
Based on the analysis of the change of SINRof users in cellrange extension region the capacity of these users with andwithout cell range extension is further analyzed Suppose thecapacity of user 119906 before and after applying CSO is 119862(119898)CRE(119906)
and 119862(119901)CRE(119906) respectively
119862(119898)
CRE (119906) = (1 minus 120588119878(119898)(119906)) log2 (1 + 120574(119898)
119906CRE) (12a)
119862(119901)
CRE (119906) = (1 minus 120588119878(119901)(119906)) log2 (1 + 120574(119901)
119906CRE) (12b)
Comparing (12a) and (12b) it can be seen that althoughthe SINR is expected to be reduced (120574(119901)
119906CRE le 120574(119898)
119906CRE) forusers in CRE region after range extension the capacity ofthese users depends on the cell load of their serving cells aswell If the cell load in the picocell is below the cell load in
0 5 10 15 20 250
10
20
30
40
50
60
70
80
90
100
CDF
()
SINR of users in CRE region
SINR (dB)
wo CREw CRE
minus15 minus10 minus5
Figure 2 SINR of users in range extension area
the macrocell by certain extent although there is an expectedSINR loss the capacity of users in the CRE region can stillexpect an increase
119862(119901)
CRE (119906) gt 119862(119898)
CRE (119906) (13a)
(1minus120588119878(119901)(119906)) log2(1+120574(119901)
119906CRE) gt (1minus120588119878(119898)(119906)) log2 (1+120574(119898)
119906CRE)
(13b)
Therefore we can derive the theorem below to show thecondition that ensures a gain in capacity of users in CREregion
Theorem 1 The capacity of users in region of CRE can beincreased if and only if the cell load in the original servingmacrocell before CRE satisfies the inequality as below
120588119878(119898)(119906)gt 1 minus (1 minus 120588
119878(119901)(119906))
log2(1 + 120574
(119901)
119906CRE)
log2(1 + 120574
(119898)
119906CRE) (14)
Comparing the role of SINR and cell load in capacityit can be seen that the cell load linearly contributes tothe capacity while SINR contributes in logarithmic scaleTherefore the contribution from cell load plays an evenmoreimportant role in capacity than SINR does and needs to betreated seriously
To summarize on one hand for users in the CREregion usually the SINR decreases due to the suboptimalreceived signal power and high interference from macrocellon the other hand CRE may also bring benefit in terms ofavailable resource (eg frequency time and power) to theseusers More resource to a user means for example moreopportunity to be scheduled by the networkwhich eventuallycontributes to the user capacity
International Journal of Antennas and Propagation 5
Suppose the total capacity of users in CRE region is119862CREand the number of users in the CRE region is119873
119877 we have
119862CRE =
119873119877
sum
119906=1
119862CRE (119906) (15)
222 Capacity Analysis for Users inMacrocell In this sectionthe capacity of users who are served by the macrocellconsistently before and after range extension is analyzed Forthese users although the serving cell is unchanged the SINRand capacity are impacted anyway by range extension
Suppose the SINR of a user 119906 in this case is representedby 120574119906119872
120574119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(16)
By applying range extension more users are connectedto the picocell and the cell load in the picocell is expectedto increase To facilitate the analysis the interference frompicocell is separated from the total interference in (16)Denoting the picocell as119901 the SINRof user119906 can be rewrittenas
120574119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 = 119901
120588119888119875119888119871119888( 119902119906) + 120588119901119875119901119871119901( 119902119906)
(17)
Denote the SINR of user 119906 after CRE as 1205741015840119906119872
Afterapplying CRE the received signal power from the servingmacrocell can be assumed to be unchanged if we assumethe userrsquos position does not change significantly during thetime of handover executionThe interference from other cellsexcept for the picocell can also be assumed to be unchangedfor simplicity However with CRE the load in the picocell1205881015840
119901 is expected to increase (1205881015840
119901gt 120588119901) Consequently the
interference from picocell will increase eventually The SINRof user 119906 after CRE is then
1205741015840
119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 = 119901
120588119888119875119888119871119888( 119902119906) + 1205881015840
119901119875119901119871119901( 119902119906)
(18)
Therefore the SINR of users served by the macrocell willdecrease due to the increased interference from picocell
1205741015840
119906119872lt 120574119906119872 (19)
Denoting the macrocell as119898 and the capacity of users inmacrocell as119862
119872(119906) the capacity is determined by both SINR
of these users and the load situation in the macrocell
119862119872(119906) = (1 minus 120588
119898) log2(1 + 120574
119906119872) (20)
Denoting the capacity of user 119906 after CRE as 1198621015840119872(119906) we
have
1198621015840
119872(119906) = (1 minus 120588
1015840
119898) log2(1 + 120574
1015840
119906119872) (21)
It is desirable that with CRE the capacity of users stay inmacrocell could increase
1198621015840
119872(119906) gt 119862
119872(119906) (22a)
(1 minus 1205881015840
119898) log2(1 + 120574
1015840
119906119872) gt (1 minus 120588
119898) log2(1 + 120574
119906119872) (22b)
Therefore we can derive the theorem as below to show thecondition that ensures a gain in capacity with CRE for usersstay in macrocell
Theorem2 Thecapacity of users inmacrocell can be increasedif and only if the cell load in themacrocell after CRE satisfies theinequality as follows
1205881015840
119898lt 1 minus (1 minus 120588
119898)
log2(1 + 120574
119906119872)
log2(1 + 120574
1015840
119906119872)
(23)
Theorem 2 shows that only when the cell load of macro-cell decreases low enough to compensate for the loss in SINRthe capacity of these users can increase
More specifically suppose the number of users in themacrocell is 119873
119872 the number of scheduled resource blocks
for each user 119906 is NRB119906 and the total number of available
resource blocks in the cell is NRBavail119898 then the macrocellload 120588
119898is
120588119898=
sum119873119872
119906=1NRB119906
NRBavail119898 (24)
Therefore the capacity of user 119906 is
119862119872(119906) = (1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872) (25)
The number of scheduled resource blocks is usually afunction of SINR Considering a typical traffic type that isnamed as constant bit rate (CBR) service as the example
NRB119906=
119863119906
119877 (120574119906119872) sdot BW (26)
where BW is the bandwidth of one PRB (eg in LTE systemBW is 180 kHz) and 119863
119906is the demanded data rate of user 119906
As shown in (4) 119877(120574) is the achievable data rate that can beestimated according to the Shannon capacity Therefore thecapacity of users stay in macrocell is actually a function ofboth SINRand the number of users inmacrocell Consideringthat the served user number linearly contributes to thecapacity while SINR contributes in logarithmic scale fromthe capacity of macrocell point of view it is desirable tohandover more users out of macrocell However from theoverall capacity point of view it is necessary to consider thecapacity change of both users stay in macrocell and usersstay in CRE region as well as users in picocell which we willanalyze in the following section
6 International Journal of Antennas and Propagation
Suppose the total capacity of users in macrocell is119862119872 we
have
119862119872=
119873119872
sum
119906=1
119862119872(119906)
=
119873119872
sum
119906=1
(1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872)
(27)
223 Capacity Analysis for Users in Picocell The capacity ofusers served by picocell both before and after applying CRE isanalyzed similarly For these users although the serving cellis unchanged the SINR and capacity are eventually impactedby range extension
Suppose the SINR of a user 119906 in this case is presented by120574119906119875
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(28)
To facilitate the analysis the macrocell interference isseparated from the total interference in (28) Denoting themacrocell as119898 the SINR of user 119906 can be rewritten as
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 120588119898119875119898119871119898( 119902119906)
(29)
Denote the SINR of user 119906 after CRE as 1205741015840119906119875
Similarto the analysis for users in macrocell after applying CREthe received signal power from the serving picocell can beassumed to be unchanged and the interference from othercells except for the macrocell can also be assumed to beunchanged for simplicity However with CRE the load in themacrocell denoted by 1205881015840
119898 is changed and usually the load in
macrocell is expected to be reduced (1205881015840119898lt 120588119898) Consequently
the interference from macrocell will decrease The SINR ofpico user u after CRE is then
1205741015840
119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 1205881015840
119898119875119898119871119898( 119902119906)
(30)
Therefore the SINR of users served by picocell increasesdue to the reduced interference as
1205741015840
119906119875gt 120574119906119875 (31)
Denoting the picocell as 119901 and the capacity of user119906 in picocell with and without CRE as 119862
119901(119906) and 1198621015840
119901(119906)
respectively then
119862119875(119906) = (1 minus 120588
119875) log2(1 + 120574
119906119875) (32)
1198621015840
119875(119906) = (1 minus 120588
1015840
119875) log2(1 + 120574
1015840
119906119875) (33)
To achieve capacity gain for users in picocell that is
1198621015840
119875(119906) gt 119862
119875(119906) (34a)
(1 minus 1205881015840
119875) log2(1 + 120574
1015840
119906119875) gt (1 minus 120588
119875) log2(1 + 120574
119906119875) (34b)
Therefore we derive the following theorem which statesthe conditions for users in picocell to achieve capacity gain byCRE
Theorem 3 The capacity of users in picocell can be increasedif and only if the cell load in the picocell after CRE satisfies theinequality as follows
1205881015840
119901lt 1 minus (1 minus 120588
119875)
log2(1 + 120574
119906119875)
log2(1 + 120574
1015840
119906119875)
(35)
FromTheorem 3 it can be concluded that if the increasedcell load in picocell does not exceed the limit as shown ininequality (35) the capacity of users in picocell can increaseOtherwise the capacity of the pico users would decrease
Suppose the number of users in the picocell is 119873119875 the
number of scheduled resource blocks for each user 119906 is NRB119906
and the total number of available resource blocks in thepicocell is NRBavail119901 The picocell load 120588
119901is then
120588119901=
sum119873119875
119906=1NRB119906
NRBavail119901 (36)
And the picocell load after CRE is
1205881015840
119901=
sum119873119875+119873119877
119906=1NRB1015840119906
NRBavail119901 (37)
where 119873119877is the number of users in CRE region and NRB1015840
119906
is the number of resource blocks of user 119906 served by pico Asshown in (26) the number of resource blocks of user119906usuallydepends on its SINR According to the analysis typically theSINR of users in CRE region would decrease while the SINRof users in picocell would increase which results in eitherincreased NRB or decreased NRB However considering theincreased number of served users119873
119877 the cell load in picocell
usually gets difficult to satisfy the inequality of (35) andresults in a decreased capacity
Suppose the total capacity of users in picocell is 119862119875 we
have
119862119875=
119873119875
sum
119906=1
119862119875(119906)
=
119873119875
sum
119906=1
(1 minus
sum119873119875
119906=1NRB119906
NRBavail119875) log2(1 + 120574
119906119875)
(38)
224 Overall Capacity Summarizing capacity of users inmacrocell in picocell and in the CRE region we can get theoverall capacity as
119862total = 119862119872 + 119862119875 + 119862CRE (39)
International Journal of Antennas and Propagation 7
From the analysis above it can be observed that withCRE the overall capacity not only depends on the change ofSINRs of users in macrocell in picocell and in CRE regionbut also depends on the load situation and change in bothmacrocell and picocell CRE may not always give benefit tothe overall capacity in the system Considering an examplethat if the traffic load in macrocell is low and there areenough radio resources that can be allocated to the servedusers the throughput of these users is consequently limited bytheir channel conditions that is SINRs In this case forcingthese users into the picocell by CRE can further damage theSINRs and consequently damage the usersrsquo throughput Onthe other hand if the traffic load in macrocell is high and theusersrsquo throughputs are limited by the available radio resourcehanding over these users to LPN by CRE can improve theusersrsquo throughputs
To get the optimal overall capacity in the next section anadaptive algorithm to optimize theCSO setting via predictionof overall capacity for different CSO values is proposed
3 Capacity Optimization Algorithm withAdaptive Cell Range Control
In this section an adaptive algorithm is proposed to optimizethe overall capacity by adjusting the CSO setting In thisalgorithm the CSO setting is updated periodically At eachupdate instance the current CSO value is updated by a newvalue if the predicted overall capacity can be increased by thisnew attempting valueThe attemptingCSO setting is obtainedby increasing or decreasing a step to the current CSO Thenew CSO value which gives the best overall capacity is thenselected The algorithm is described step by step as in below
31 Step 1 User Grouping with Attempting CSO SettingsSuppose CSO updating step is ΔCSO At each update instance119905 CSO value is updated based on current CSO value CSO
119905minus1
whichwas set at time interval 119905minus1The attemptingCSO valueCSO119905att is obtained by either increasing or decreasingCSO119905minus1
by the updating step ΔCSO as
CSO119905att = CSO119905inc = CSO119905minus1 + ΔCSO (40a)
CSO119905att = CSO119905dec = CSO119905minus1 minus ΔCSO (40b)
Based on the RSRP measurements from both macrocelland picocell for the attempted CSO setting all of the userscan be divided into three groups
Group 1Macro users These users stay in the macrocell bothwith the previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att le RSRP119898
119906
(41)
Group 2 Pico usersThese users stay in the picocell both withthe previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1gt RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(42)
Group 3 Users in CRE region These users will change theirserving cell with new CSO setting Take the case that CSO isincreased as the following example
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(43)
On the basis of the user grouping the capacities of usersin different groups are predicted respectively and the overallcapacity is predicted consequently
32 Step 2 Capacity Prediction with Attempting CSO SettingsIn step 2 the capacities of users in different groups arepredicted respectively according to the predicted SINR andcell load More details can be found below
321 Step 21 Capacity Prediction for Users in CRE RegionTo estimate the new capacity of users in CRE region we needto estimate the new SINR and predict the new cell load inpicocell and in macrocell
Taking the case that the CSO is increased by a step asexample to estimate the new SINR in picocell 120574(119901)
119906CRE we canuse
120574(119901)
119906CRE =RSRP(119901)
(RSRP(119898)120574(119898)119906CRE) + RSRP
(119898)minus RSRP(119901)
(44)
To estimate the new picocell load we need to estimate therequired number of RBs according to the new SINR For CBRservice this can be estimated similarly to (26) by
119873119906=
119863119906
119877 (120574(119901)
119906CRE) sdot BW (45)
For services other than CBR service the new requirednumber of RBs can be estimated according to the relative rela-tionship between the previous SINR in macrocell previousrequired number of RBs inmacrocell and the predicted SINRin picocell
119873119906
new= 119873119906
oldsdot
119877 (120574(119898)
119906CRE)
119877 (120574(119901)
119906CRE) (46)
Based on the prediction of the required number of RBsthe new picocell load can be estimated by assuming therequired number of RBs is unchanged for other users inpicocell
120588119901
new=
sum119873119875
119906=1119873119906+ sum119873119877
119906=1119873119906
new
NRBavail119901
(47)
8 International Journal of Antennas and Propagation
Consequently the new capacity of users newly added toCRE region can be estimated according to the analysis inSection 2
322 Step 22 Capacity Prediction for Macro Users and PicoUsers Formacro users to predict the SINR of users with newCSO the picocell load with new CSO needs to be estimatedThis can be achieved by method described in Step 21
The new cell load in macrocell can be simply estimatedby the new number of served users and assuming that therequired numbers of RBs for macro users are unchanged
The capacity of users in picocell can be estimated simi-larly
33 Optimal CSO Selection and Updating The predictednew capacities of users in different groups are summedtogether to get the overall predicted capacity The predictedoverall capacity is then compared with current capacity If theestimated new capacity is larger than the current capacity theCSO is updated
If 119862 (CSO119905att) gt 119862 (CSO119905minus1) CSO119905 = CSO119905att (48)
Otherwise CSO keeps unchangedIf the CSO is decreased by a step the overall capacity
can be estimated similarly by reversing the calculation shownabove If the new CSO value gives better predicted perfor-mance the CSO value is updated
With the proposed adaptive algorithm the CSO value isadjusted according to the predicted overall capacity in bothmacrocell and picocell By selecting the most suitable CSOvalue the overall capacity is expected to be optimized
4 Numerical Results
The proposed adaptive CSO updating algorithm is evaluatedvia system-level simulations In this section the simulationresults are presented and discussed The simulation assump-tions and parameter settings are illustrated in Table 1
Two typical scenarios are tested in the evaluation Inscenario 1 the users are uniformly distributed in the entiresimulated area LPNs are deployed randomly In scenario 2LPNs are deployed in hotspot areas and the hotspot areastake nearly 60of the usersThe base station deployment anduser distribution in two scenarios are illustrated in Figures 3and 4 To have a clear view only part of the simulated areas isillustrated
The percentages of users belonging to macrocell or LPNsin the two scenarios are illustrated in Figure 5 with differentCSO settings It can be observed that when CSO settingincreases the users belonging to LPNs increase It can beobserved as well that the percentages of users belonging toLPNs in hotspot scenario are much higher than in uniformscenario
Figure 6 shows the change of average cell load withdifferent CSO settings in different scenarios In uniformscenario most of the users are in macrocells and the loadin the macrocell is consequently quite high the cell load inpicocell increases with the increase in the CSO setting In
0
1
2
3
4
5
6
911
15 17
Figure 3 Cell layout and user distribution in uniform scenario
0
1
2
3
4
5
6
911
15 17
Figure 4 Cell layout and user distribution in hotspot scenario
hotspot scenario the average cell load inmacrocells decreaseswith the increase in CSO setting while the average cell load inLPNs increases further with the increased CSO setting
In Figure 7 the average sector throughput gain withdifferent CSO settings is compared with the proposed adap-tive algorithm for uniform scenario The average sectorthroughput with CSO = 0 dB is used as reference to obtainthe average sector throughput gain It can be observed that inthis scenario a higherCSO setting is desirable to offloadmoreusers to LPNs The proposed adaptive algorithm can achievethe similar throughput to the throughput of the optimal CSOsetting
In Figure 8 the average sector throughput gain with dif-ferent CSO settings is compared with the proposed adaptivealgorithm in hotspot scenarioThe average sector throughput
International Journal of Antennas and Propagation 9
0001020304050607080910
Uniform-macroUniform-pico
Hotspot-picoHotspot-macro
CSO 15CSO 10CSO 5CSO 0
Con
nect
ion
ratio
Figure 5 Connection ratio with different CSOs
0 2 4 6 8 10
05
06
07
08
09
10
Uniform-macroUniform-pico
Hotspot-macroHotspot-pico
CSO (dB)
Nor
mal
ized
cell
load
Figure 6 Cell load with different CSOs
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 7 Average sector throughput gain in uniform scenario
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 8 Average sector throughput gain in hotspot scenario
with CSO = 0 dB is used as reference to obtain the averagesector throughput gain In this scenario CSO of 10 dB isdesirable to balance the offloading gain and SINR loss Theproposed algorithm achieves the similar performance to theoptimal CSO setting
In different scenarios the optimal CSO setting is differ-ent However the proposed adaptive algorithm can achievesimilar performance with the optimal CSO setting in bothscenarios
5 Conclusion
As an attractivemeans of expandingmobile network capacityheterogeneous network has been included in LTE-advancedCell range extension is an important feature in heterogeneousnetwork to improve the utilization of the resource of lowpower nodes However the users in the cell range extensionarea usually suffer from high interference from macrocellsTo avoid the unnecessary interference it is desirable tocarefully consider the optimal cell selection offset value Inthis paper the capacity of users in macrocell and small cellsis analyzed thoroughly The conditions to improve capacityvia cell range extension are derived respectively for usersboth in macrocell and in small cell Based on the analysis anadaptive small cell coverage control algorithm is proposedWith the proposed adaptive algorithm the cell selectionoffset is updated periodically based on the prediction of theoverall system capacityThe algorithm is evaluated by system-level simulations and the results indicate that with theproposed algorithm a nearly optimal system performancecan be achieved in all tested load cases
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
10 International Journal of Antennas and Propagation
Acknowledgments
This work was supported by the State Major Science andTechnology Special Projects (Grant no 2013ZX03001026-001) and the Fundamental Research Funds for the CentralUniversities (Grant no 2014RC0107)
References
[1] A Damnjanovic J Montojo Y Wei et al ldquoA survey on 3GPPheterogeneous networksrdquo IEEE Wireless Communications vol18 no 3 pp 10ndash21 2011
[2] E Dahlman S Parkvall and J Skold 4G LTELTE-Advanced forMobile Broadband Academic Press 2011
[3] A Ghosh R Ratasuk B Mondal N Mangalvedhe and TThomas ldquoLTE-advanced next-generation wireless broadbandtechnologyrdquo IEEE Wireless Communications vol 17 no 3 pp10ndash22 2010
[4] X Zhang XGuW Li L Zhang J Shen andYWan ldquoThe studyof indoor and field trials on 2times8MIMOarchitecture in TD-LTEnetworkrdquo International Journal of Antennas and Propagationsvol 2013 Article ID 181579 9 pages 2013
[5] G Yuan X Zhang WWang and Y Yang ldquoCarrier aggregationfor LTE-advancedmobile communication systemsrdquo IEEE Com-munications Magazine vol 48 no 2 pp 88ndash93 2010
[6] R Irmer H Droste P Marsch et al ldquoCoordinated multipointconcepts performance and field trial resultsrdquo IEEE Communi-cations Magazine vol 49 no 2 pp 102ndash111 2011
[7] I Siomina and Y Di ldquoLoad balancing in heterogeneous LTErange optimization via cell offset and load-coupling character-izationrdquo in Proceedings of the IEEE International Conference onCommunications pp 1357ndash1361 2012
[8] The 3rd Generation Partnership Project (3GPP) ldquoFeasibilitystudy for further advancements for E-UTRA (LTEAdvanced)(Release 10)rdquo Tech Rep TR 36912 2011
[9] The 3rd Generation Partnership Project(3GPP) ldquoSystem per-formance of heterogeneous networks with range expansionrdquoTech Rep R1-101203 Samsung 2010
[10] I Guvenc ldquoCapacity and fairness analysis of heterogeneous net-works with range expansion and interference coordinationrdquoIEEECommunications Letters vol 15 no 10 pp 1084ndash1087 2011
[11] H-S Jo Y J Sang P Xia and J G Andrews ldquoHeterogeneouscellular networks with flexible cell association a comprehensivedownlink SINR analysisrdquo IEEE Transactions on Wireless Com-munications vol 11 no 10 pp 3484ndash3495 2012
[12] ldquoSelf-optimisationand self-configuration in wireless networksrdquoSOCRATES European Research Project httpwwwfp7-socratesorg
[13] The 3rd Generation Partnership Project(3GPP) ldquoSelf-config-uring and self-optimizing network use cases and solutionsrdquoTech Rep TR 36902 2009
[14] ldquoUse cases related to Self Organizing network Overall descrip-tionrdquoNext generationMobileNetworks httpwwwngmnorg
[15] X Chu and D Lopez-Perez Heterogeneous Cellular NetworksTheory Simulation and Deployment Cambridge UniversityPress 2013
[16] P Tian H Tian J Zhu L Chen and X She ldquoAn adaptive biasconfiguration strategy for range extension in LTE-Advancedheterogeneous networksrdquo inProceedings of the IET InternationalConference on Communication Technology and Application(ICCTA rsquo11) pp 336ndash340 2011
[17] K Kikuchi and H Otsuka ldquoProposal of adaptive control CREin heterogeneous networksrdquo in Proceedings of the IEEE Inter-national Symposium on Personal Indoor and Mobile RadioCommunications (PIMRC rsquo12) pp 910ndash914 2012
[18] I Viering M Dottling and A Lobinger ldquoA mathematical per-spective of self-optimizing wireless networksrdquo in Proceedings ofthe IEEE International Conference on Communications (ICCrsquo09) pp 1ndash6 June 2009
[19] A Lobinger S Stefanski T Jansen and I Balan ldquoLoad balanc-ing in downlink LTE self-optimizing networksrdquo in Proceedingsof the IEEE 71st Vehicular Technology Conference (VTC rsquo10) pp1ndash5 May 2010
[20] ldquoGuidelines for evaluation of radio transmission technologiesfor IMT-2000rdquo Recommendation ITU-R M1225 1997
[21] The 3rd Generation Partnership Project (3GPP) ldquoFurtheradvancements for E-UTRAN physical layer aspects (Release9)rdquo Tech Rep TS 36814 2010
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Antennas and Propagation 5
Suppose the total capacity of users in CRE region is119862CREand the number of users in the CRE region is119873
119877 we have
119862CRE =
119873119877
sum
119906=1
119862CRE (119906) (15)
222 Capacity Analysis for Users inMacrocell In this sectionthe capacity of users who are served by the macrocellconsistently before and after range extension is analyzed Forthese users although the serving cell is unchanged the SINRand capacity are impacted anyway by range extension
Suppose the SINR of a user 119906 in this case is representedby 120574119906119872
120574119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(16)
By applying range extension more users are connectedto the picocell and the cell load in the picocell is expectedto increase To facilitate the analysis the interference frompicocell is separated from the total interference in (16)Denoting the picocell as119901 the SINRof user119906 can be rewrittenas
120574119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 = 119901
120588119888119875119888119871119888( 119902119906) + 120588119901119875119901119871119901( 119902119906)
(17)
Denote the SINR of user 119906 after CRE as 1205741015840119906119872
Afterapplying CRE the received signal power from the servingmacrocell can be assumed to be unchanged if we assumethe userrsquos position does not change significantly during thetime of handover executionThe interference from other cellsexcept for the picocell can also be assumed to be unchangedfor simplicity However with CRE the load in the picocell1205881015840
119901 is expected to increase (1205881015840
119901gt 120588119901) Consequently the
interference from picocell will increase eventually The SINRof user 119906 after CRE is then
1205741015840
119906119872=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 = 119901
120588119888119875119888119871119888( 119902119906) + 1205881015840
119901119875119901119871119901( 119902119906)
(18)
Therefore the SINR of users served by the macrocell willdecrease due to the increased interference from picocell
1205741015840
119906119872lt 120574119906119872 (19)
Denoting the macrocell as119898 and the capacity of users inmacrocell as119862
119872(119906) the capacity is determined by both SINR
of these users and the load situation in the macrocell
119862119872(119906) = (1 minus 120588
119898) log2(1 + 120574
119906119872) (20)
Denoting the capacity of user 119906 after CRE as 1198621015840119872(119906) we
have
1198621015840
119872(119906) = (1 minus 120588
1015840
119898) log2(1 + 120574
1015840
119906119872) (21)
It is desirable that with CRE the capacity of users stay inmacrocell could increase
1198621015840
119872(119906) gt 119862
119872(119906) (22a)
(1 minus 1205881015840
119898) log2(1 + 120574
1015840
119906119872) gt (1 minus 120588
119898) log2(1 + 120574
119906119872) (22b)
Therefore we can derive the theorem as below to show thecondition that ensures a gain in capacity with CRE for usersstay in macrocell
Theorem2 Thecapacity of users inmacrocell can be increasedif and only if the cell load in themacrocell after CRE satisfies theinequality as follows
1205881015840
119898lt 1 minus (1 minus 120588
119898)
log2(1 + 120574
119906119872)
log2(1 + 120574
1015840
119906119872)
(23)
Theorem 2 shows that only when the cell load of macro-cell decreases low enough to compensate for the loss in SINRthe capacity of these users can increase
More specifically suppose the number of users in themacrocell is 119873
119872 the number of scheduled resource blocks
for each user 119906 is NRB119906 and the total number of available
resource blocks in the cell is NRBavail119898 then the macrocellload 120588
119898is
120588119898=
sum119873119872
119906=1NRB119906
NRBavail119898 (24)
Therefore the capacity of user 119906 is
119862119872(119906) = (1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872) (25)
The number of scheduled resource blocks is usually afunction of SINR Considering a typical traffic type that isnamed as constant bit rate (CBR) service as the example
NRB119906=
119863119906
119877 (120574119906119872) sdot BW (26)
where BW is the bandwidth of one PRB (eg in LTE systemBW is 180 kHz) and 119863
119906is the demanded data rate of user 119906
As shown in (4) 119877(120574) is the achievable data rate that can beestimated according to the Shannon capacity Therefore thecapacity of users stay in macrocell is actually a function ofboth SINRand the number of users inmacrocell Consideringthat the served user number linearly contributes to thecapacity while SINR contributes in logarithmic scale fromthe capacity of macrocell point of view it is desirable tohandover more users out of macrocell However from theoverall capacity point of view it is necessary to consider thecapacity change of both users stay in macrocell and usersstay in CRE region as well as users in picocell which we willanalyze in the following section
6 International Journal of Antennas and Propagation
Suppose the total capacity of users in macrocell is119862119872 we
have
119862119872=
119873119872
sum
119906=1
119862119872(119906)
=
119873119872
sum
119906=1
(1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872)
(27)
223 Capacity Analysis for Users in Picocell The capacity ofusers served by picocell both before and after applying CRE isanalyzed similarly For these users although the serving cellis unchanged the SINR and capacity are eventually impactedby range extension
Suppose the SINR of a user 119906 in this case is presented by120574119906119875
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(28)
To facilitate the analysis the macrocell interference isseparated from the total interference in (28) Denoting themacrocell as119898 the SINR of user 119906 can be rewritten as
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 120588119898119875119898119871119898( 119902119906)
(29)
Denote the SINR of user 119906 after CRE as 1205741015840119906119875
Similarto the analysis for users in macrocell after applying CREthe received signal power from the serving picocell can beassumed to be unchanged and the interference from othercells except for the macrocell can also be assumed to beunchanged for simplicity However with CRE the load in themacrocell denoted by 1205881015840
119898 is changed and usually the load in
macrocell is expected to be reduced (1205881015840119898lt 120588119898) Consequently
the interference from macrocell will decrease The SINR ofpico user u after CRE is then
1205741015840
119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 1205881015840
119898119875119898119871119898( 119902119906)
(30)
Therefore the SINR of users served by picocell increasesdue to the reduced interference as
1205741015840
119906119875gt 120574119906119875 (31)
Denoting the picocell as 119901 and the capacity of user119906 in picocell with and without CRE as 119862
119901(119906) and 1198621015840
119901(119906)
respectively then
119862119875(119906) = (1 minus 120588
119875) log2(1 + 120574
119906119875) (32)
1198621015840
119875(119906) = (1 minus 120588
1015840
119875) log2(1 + 120574
1015840
119906119875) (33)
To achieve capacity gain for users in picocell that is
1198621015840
119875(119906) gt 119862
119875(119906) (34a)
(1 minus 1205881015840
119875) log2(1 + 120574
1015840
119906119875) gt (1 minus 120588
119875) log2(1 + 120574
119906119875) (34b)
Therefore we derive the following theorem which statesthe conditions for users in picocell to achieve capacity gain byCRE
Theorem 3 The capacity of users in picocell can be increasedif and only if the cell load in the picocell after CRE satisfies theinequality as follows
1205881015840
119901lt 1 minus (1 minus 120588
119875)
log2(1 + 120574
119906119875)
log2(1 + 120574
1015840
119906119875)
(35)
FromTheorem 3 it can be concluded that if the increasedcell load in picocell does not exceed the limit as shown ininequality (35) the capacity of users in picocell can increaseOtherwise the capacity of the pico users would decrease
Suppose the number of users in the picocell is 119873119875 the
number of scheduled resource blocks for each user 119906 is NRB119906
and the total number of available resource blocks in thepicocell is NRBavail119901 The picocell load 120588
119901is then
120588119901=
sum119873119875
119906=1NRB119906
NRBavail119901 (36)
And the picocell load after CRE is
1205881015840
119901=
sum119873119875+119873119877
119906=1NRB1015840119906
NRBavail119901 (37)
where 119873119877is the number of users in CRE region and NRB1015840
119906
is the number of resource blocks of user 119906 served by pico Asshown in (26) the number of resource blocks of user119906usuallydepends on its SINR According to the analysis typically theSINR of users in CRE region would decrease while the SINRof users in picocell would increase which results in eitherincreased NRB or decreased NRB However considering theincreased number of served users119873
119877 the cell load in picocell
usually gets difficult to satisfy the inequality of (35) andresults in a decreased capacity
Suppose the total capacity of users in picocell is 119862119875 we
have
119862119875=
119873119875
sum
119906=1
119862119875(119906)
=
119873119875
sum
119906=1
(1 minus
sum119873119875
119906=1NRB119906
NRBavail119875) log2(1 + 120574
119906119875)
(38)
224 Overall Capacity Summarizing capacity of users inmacrocell in picocell and in the CRE region we can get theoverall capacity as
119862total = 119862119872 + 119862119875 + 119862CRE (39)
International Journal of Antennas and Propagation 7
From the analysis above it can be observed that withCRE the overall capacity not only depends on the change ofSINRs of users in macrocell in picocell and in CRE regionbut also depends on the load situation and change in bothmacrocell and picocell CRE may not always give benefit tothe overall capacity in the system Considering an examplethat if the traffic load in macrocell is low and there areenough radio resources that can be allocated to the servedusers the throughput of these users is consequently limited bytheir channel conditions that is SINRs In this case forcingthese users into the picocell by CRE can further damage theSINRs and consequently damage the usersrsquo throughput Onthe other hand if the traffic load in macrocell is high and theusersrsquo throughputs are limited by the available radio resourcehanding over these users to LPN by CRE can improve theusersrsquo throughputs
To get the optimal overall capacity in the next section anadaptive algorithm to optimize theCSO setting via predictionof overall capacity for different CSO values is proposed
3 Capacity Optimization Algorithm withAdaptive Cell Range Control
In this section an adaptive algorithm is proposed to optimizethe overall capacity by adjusting the CSO setting In thisalgorithm the CSO setting is updated periodically At eachupdate instance the current CSO value is updated by a newvalue if the predicted overall capacity can be increased by thisnew attempting valueThe attemptingCSO setting is obtainedby increasing or decreasing a step to the current CSO Thenew CSO value which gives the best overall capacity is thenselected The algorithm is described step by step as in below
31 Step 1 User Grouping with Attempting CSO SettingsSuppose CSO updating step is ΔCSO At each update instance119905 CSO value is updated based on current CSO value CSO
119905minus1
whichwas set at time interval 119905minus1The attemptingCSO valueCSO119905att is obtained by either increasing or decreasingCSO119905minus1
by the updating step ΔCSO as
CSO119905att = CSO119905inc = CSO119905minus1 + ΔCSO (40a)
CSO119905att = CSO119905dec = CSO119905minus1 minus ΔCSO (40b)
Based on the RSRP measurements from both macrocelland picocell for the attempted CSO setting all of the userscan be divided into three groups
Group 1Macro users These users stay in the macrocell bothwith the previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att le RSRP119898
119906
(41)
Group 2 Pico usersThese users stay in the picocell both withthe previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1gt RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(42)
Group 3 Users in CRE region These users will change theirserving cell with new CSO setting Take the case that CSO isincreased as the following example
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(43)
On the basis of the user grouping the capacities of usersin different groups are predicted respectively and the overallcapacity is predicted consequently
32 Step 2 Capacity Prediction with Attempting CSO SettingsIn step 2 the capacities of users in different groups arepredicted respectively according to the predicted SINR andcell load More details can be found below
321 Step 21 Capacity Prediction for Users in CRE RegionTo estimate the new capacity of users in CRE region we needto estimate the new SINR and predict the new cell load inpicocell and in macrocell
Taking the case that the CSO is increased by a step asexample to estimate the new SINR in picocell 120574(119901)
119906CRE we canuse
120574(119901)
119906CRE =RSRP(119901)
(RSRP(119898)120574(119898)119906CRE) + RSRP
(119898)minus RSRP(119901)
(44)
To estimate the new picocell load we need to estimate therequired number of RBs according to the new SINR For CBRservice this can be estimated similarly to (26) by
119873119906=
119863119906
119877 (120574(119901)
119906CRE) sdot BW (45)
For services other than CBR service the new requirednumber of RBs can be estimated according to the relative rela-tionship between the previous SINR in macrocell previousrequired number of RBs inmacrocell and the predicted SINRin picocell
119873119906
new= 119873119906
oldsdot
119877 (120574(119898)
119906CRE)
119877 (120574(119901)
119906CRE) (46)
Based on the prediction of the required number of RBsthe new picocell load can be estimated by assuming therequired number of RBs is unchanged for other users inpicocell
120588119901
new=
sum119873119875
119906=1119873119906+ sum119873119877
119906=1119873119906
new
NRBavail119901
(47)
8 International Journal of Antennas and Propagation
Consequently the new capacity of users newly added toCRE region can be estimated according to the analysis inSection 2
322 Step 22 Capacity Prediction for Macro Users and PicoUsers Formacro users to predict the SINR of users with newCSO the picocell load with new CSO needs to be estimatedThis can be achieved by method described in Step 21
The new cell load in macrocell can be simply estimatedby the new number of served users and assuming that therequired numbers of RBs for macro users are unchanged
The capacity of users in picocell can be estimated simi-larly
33 Optimal CSO Selection and Updating The predictednew capacities of users in different groups are summedtogether to get the overall predicted capacity The predictedoverall capacity is then compared with current capacity If theestimated new capacity is larger than the current capacity theCSO is updated
If 119862 (CSO119905att) gt 119862 (CSO119905minus1) CSO119905 = CSO119905att (48)
Otherwise CSO keeps unchangedIf the CSO is decreased by a step the overall capacity
can be estimated similarly by reversing the calculation shownabove If the new CSO value gives better predicted perfor-mance the CSO value is updated
With the proposed adaptive algorithm the CSO value isadjusted according to the predicted overall capacity in bothmacrocell and picocell By selecting the most suitable CSOvalue the overall capacity is expected to be optimized
4 Numerical Results
The proposed adaptive CSO updating algorithm is evaluatedvia system-level simulations In this section the simulationresults are presented and discussed The simulation assump-tions and parameter settings are illustrated in Table 1
Two typical scenarios are tested in the evaluation Inscenario 1 the users are uniformly distributed in the entiresimulated area LPNs are deployed randomly In scenario 2LPNs are deployed in hotspot areas and the hotspot areastake nearly 60of the usersThe base station deployment anduser distribution in two scenarios are illustrated in Figures 3and 4 To have a clear view only part of the simulated areas isillustrated
The percentages of users belonging to macrocell or LPNsin the two scenarios are illustrated in Figure 5 with differentCSO settings It can be observed that when CSO settingincreases the users belonging to LPNs increase It can beobserved as well that the percentages of users belonging toLPNs in hotspot scenario are much higher than in uniformscenario
Figure 6 shows the change of average cell load withdifferent CSO settings in different scenarios In uniformscenario most of the users are in macrocells and the loadin the macrocell is consequently quite high the cell load inpicocell increases with the increase in the CSO setting In
0
1
2
3
4
5
6
911
15 17
Figure 3 Cell layout and user distribution in uniform scenario
0
1
2
3
4
5
6
911
15 17
Figure 4 Cell layout and user distribution in hotspot scenario
hotspot scenario the average cell load inmacrocells decreaseswith the increase in CSO setting while the average cell load inLPNs increases further with the increased CSO setting
In Figure 7 the average sector throughput gain withdifferent CSO settings is compared with the proposed adap-tive algorithm for uniform scenario The average sectorthroughput with CSO = 0 dB is used as reference to obtainthe average sector throughput gain It can be observed that inthis scenario a higherCSO setting is desirable to offloadmoreusers to LPNs The proposed adaptive algorithm can achievethe similar throughput to the throughput of the optimal CSOsetting
In Figure 8 the average sector throughput gain with dif-ferent CSO settings is compared with the proposed adaptivealgorithm in hotspot scenarioThe average sector throughput
International Journal of Antennas and Propagation 9
0001020304050607080910
Uniform-macroUniform-pico
Hotspot-picoHotspot-macro
CSO 15CSO 10CSO 5CSO 0
Con
nect
ion
ratio
Figure 5 Connection ratio with different CSOs
0 2 4 6 8 10
05
06
07
08
09
10
Uniform-macroUniform-pico
Hotspot-macroHotspot-pico
CSO (dB)
Nor
mal
ized
cell
load
Figure 6 Cell load with different CSOs
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 7 Average sector throughput gain in uniform scenario
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 8 Average sector throughput gain in hotspot scenario
with CSO = 0 dB is used as reference to obtain the averagesector throughput gain In this scenario CSO of 10 dB isdesirable to balance the offloading gain and SINR loss Theproposed algorithm achieves the similar performance to theoptimal CSO setting
In different scenarios the optimal CSO setting is differ-ent However the proposed adaptive algorithm can achievesimilar performance with the optimal CSO setting in bothscenarios
5 Conclusion
As an attractivemeans of expandingmobile network capacityheterogeneous network has been included in LTE-advancedCell range extension is an important feature in heterogeneousnetwork to improve the utilization of the resource of lowpower nodes However the users in the cell range extensionarea usually suffer from high interference from macrocellsTo avoid the unnecessary interference it is desirable tocarefully consider the optimal cell selection offset value Inthis paper the capacity of users in macrocell and small cellsis analyzed thoroughly The conditions to improve capacityvia cell range extension are derived respectively for usersboth in macrocell and in small cell Based on the analysis anadaptive small cell coverage control algorithm is proposedWith the proposed adaptive algorithm the cell selectionoffset is updated periodically based on the prediction of theoverall system capacityThe algorithm is evaluated by system-level simulations and the results indicate that with theproposed algorithm a nearly optimal system performancecan be achieved in all tested load cases
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
10 International Journal of Antennas and Propagation
Acknowledgments
This work was supported by the State Major Science andTechnology Special Projects (Grant no 2013ZX03001026-001) and the Fundamental Research Funds for the CentralUniversities (Grant no 2014RC0107)
References
[1] A Damnjanovic J Montojo Y Wei et al ldquoA survey on 3GPPheterogeneous networksrdquo IEEE Wireless Communications vol18 no 3 pp 10ndash21 2011
[2] E Dahlman S Parkvall and J Skold 4G LTELTE-Advanced forMobile Broadband Academic Press 2011
[3] A Ghosh R Ratasuk B Mondal N Mangalvedhe and TThomas ldquoLTE-advanced next-generation wireless broadbandtechnologyrdquo IEEE Wireless Communications vol 17 no 3 pp10ndash22 2010
[4] X Zhang XGuW Li L Zhang J Shen andYWan ldquoThe studyof indoor and field trials on 2times8MIMOarchitecture in TD-LTEnetworkrdquo International Journal of Antennas and Propagationsvol 2013 Article ID 181579 9 pages 2013
[5] G Yuan X Zhang WWang and Y Yang ldquoCarrier aggregationfor LTE-advancedmobile communication systemsrdquo IEEE Com-munications Magazine vol 48 no 2 pp 88ndash93 2010
[6] R Irmer H Droste P Marsch et al ldquoCoordinated multipointconcepts performance and field trial resultsrdquo IEEE Communi-cations Magazine vol 49 no 2 pp 102ndash111 2011
[7] I Siomina and Y Di ldquoLoad balancing in heterogeneous LTErange optimization via cell offset and load-coupling character-izationrdquo in Proceedings of the IEEE International Conference onCommunications pp 1357ndash1361 2012
[8] The 3rd Generation Partnership Project (3GPP) ldquoFeasibilitystudy for further advancements for E-UTRA (LTEAdvanced)(Release 10)rdquo Tech Rep TR 36912 2011
[9] The 3rd Generation Partnership Project(3GPP) ldquoSystem per-formance of heterogeneous networks with range expansionrdquoTech Rep R1-101203 Samsung 2010
[10] I Guvenc ldquoCapacity and fairness analysis of heterogeneous net-works with range expansion and interference coordinationrdquoIEEECommunications Letters vol 15 no 10 pp 1084ndash1087 2011
[11] H-S Jo Y J Sang P Xia and J G Andrews ldquoHeterogeneouscellular networks with flexible cell association a comprehensivedownlink SINR analysisrdquo IEEE Transactions on Wireless Com-munications vol 11 no 10 pp 3484ndash3495 2012
[12] ldquoSelf-optimisationand self-configuration in wireless networksrdquoSOCRATES European Research Project httpwwwfp7-socratesorg
[13] The 3rd Generation Partnership Project(3GPP) ldquoSelf-config-uring and self-optimizing network use cases and solutionsrdquoTech Rep TR 36902 2009
[14] ldquoUse cases related to Self Organizing network Overall descrip-tionrdquoNext generationMobileNetworks httpwwwngmnorg
[15] X Chu and D Lopez-Perez Heterogeneous Cellular NetworksTheory Simulation and Deployment Cambridge UniversityPress 2013
[16] P Tian H Tian J Zhu L Chen and X She ldquoAn adaptive biasconfiguration strategy for range extension in LTE-Advancedheterogeneous networksrdquo inProceedings of the IET InternationalConference on Communication Technology and Application(ICCTA rsquo11) pp 336ndash340 2011
[17] K Kikuchi and H Otsuka ldquoProposal of adaptive control CREin heterogeneous networksrdquo in Proceedings of the IEEE Inter-national Symposium on Personal Indoor and Mobile RadioCommunications (PIMRC rsquo12) pp 910ndash914 2012
[18] I Viering M Dottling and A Lobinger ldquoA mathematical per-spective of self-optimizing wireless networksrdquo in Proceedings ofthe IEEE International Conference on Communications (ICCrsquo09) pp 1ndash6 June 2009
[19] A Lobinger S Stefanski T Jansen and I Balan ldquoLoad balanc-ing in downlink LTE self-optimizing networksrdquo in Proceedingsof the IEEE 71st Vehicular Technology Conference (VTC rsquo10) pp1ndash5 May 2010
[20] ldquoGuidelines for evaluation of radio transmission technologiesfor IMT-2000rdquo Recommendation ITU-R M1225 1997
[21] The 3rd Generation Partnership Project (3GPP) ldquoFurtheradvancements for E-UTRAN physical layer aspects (Release9)rdquo Tech Rep TS 36814 2010
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
6 International Journal of Antennas and Propagation
Suppose the total capacity of users in macrocell is119862119872 we
have
119862119872=
119873119872
sum
119906=1
119862119872(119906)
=
119873119872
sum
119906=1
(1 minus
sum119873119872
119906=1NRB119906
NRBavail119898) log2(1 + 120574
119906119872)
(27)
223 Capacity Analysis for Users in Picocell The capacity ofusers served by picocell both before and after applying CRE isanalyzed similarly For these users although the serving cellis unchanged the SINR and capacity are eventually impactedby range extension
Suppose the SINR of a user 119906 in this case is presented by120574119906119875
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
(28)
To facilitate the analysis the macrocell interference isseparated from the total interference in (28) Denoting themacrocell as119898 the SINR of user 119906 can be rewritten as
120574119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum119888 = 119878(119906)120588119888119875119888119871119888( 119902119906)
=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 120588119898119875119898119871119898( 119902119906)
(29)
Denote the SINR of user 119906 after CRE as 1205741015840119906119875
Similarto the analysis for users in macrocell after applying CREthe received signal power from the serving picocell can beassumed to be unchanged and the interference from othercells except for the macrocell can also be assumed to beunchanged for simplicity However with CRE the load in themacrocell denoted by 1205881015840
119898 is changed and usually the load in
macrocell is expected to be reduced (1205881015840119898lt 120588119898) Consequently
the interference from macrocell will decrease The SINR ofpico user u after CRE is then
1205741015840
119906119875=
119875119878(119906)119871119878(119906)( 119902119906)
119873 + sum 119888 = 119878(119906)
119888 =119898
120588119888119875119888119871119888( 119902119906) + 1205881015840
119898119875119898119871119898( 119902119906)
(30)
Therefore the SINR of users served by picocell increasesdue to the reduced interference as
1205741015840
119906119875gt 120574119906119875 (31)
Denoting the picocell as 119901 and the capacity of user119906 in picocell with and without CRE as 119862
119901(119906) and 1198621015840
119901(119906)
respectively then
119862119875(119906) = (1 minus 120588
119875) log2(1 + 120574
119906119875) (32)
1198621015840
119875(119906) = (1 minus 120588
1015840
119875) log2(1 + 120574
1015840
119906119875) (33)
To achieve capacity gain for users in picocell that is
1198621015840
119875(119906) gt 119862
119875(119906) (34a)
(1 minus 1205881015840
119875) log2(1 + 120574
1015840
119906119875) gt (1 minus 120588
119875) log2(1 + 120574
119906119875) (34b)
Therefore we derive the following theorem which statesthe conditions for users in picocell to achieve capacity gain byCRE
Theorem 3 The capacity of users in picocell can be increasedif and only if the cell load in the picocell after CRE satisfies theinequality as follows
1205881015840
119901lt 1 minus (1 minus 120588
119875)
log2(1 + 120574
119906119875)
log2(1 + 120574
1015840
119906119875)
(35)
FromTheorem 3 it can be concluded that if the increasedcell load in picocell does not exceed the limit as shown ininequality (35) the capacity of users in picocell can increaseOtherwise the capacity of the pico users would decrease
Suppose the number of users in the picocell is 119873119875 the
number of scheduled resource blocks for each user 119906 is NRB119906
and the total number of available resource blocks in thepicocell is NRBavail119901 The picocell load 120588
119901is then
120588119901=
sum119873119875
119906=1NRB119906
NRBavail119901 (36)
And the picocell load after CRE is
1205881015840
119901=
sum119873119875+119873119877
119906=1NRB1015840119906
NRBavail119901 (37)
where 119873119877is the number of users in CRE region and NRB1015840
119906
is the number of resource blocks of user 119906 served by pico Asshown in (26) the number of resource blocks of user119906usuallydepends on its SINR According to the analysis typically theSINR of users in CRE region would decrease while the SINRof users in picocell would increase which results in eitherincreased NRB or decreased NRB However considering theincreased number of served users119873
119877 the cell load in picocell
usually gets difficult to satisfy the inequality of (35) andresults in a decreased capacity
Suppose the total capacity of users in picocell is 119862119875 we
have
119862119875=
119873119875
sum
119906=1
119862119875(119906)
=
119873119875
sum
119906=1
(1 minus
sum119873119875
119906=1NRB119906
NRBavail119875) log2(1 + 120574
119906119875)
(38)
224 Overall Capacity Summarizing capacity of users inmacrocell in picocell and in the CRE region we can get theoverall capacity as
119862total = 119862119872 + 119862119875 + 119862CRE (39)
International Journal of Antennas and Propagation 7
From the analysis above it can be observed that withCRE the overall capacity not only depends on the change ofSINRs of users in macrocell in picocell and in CRE regionbut also depends on the load situation and change in bothmacrocell and picocell CRE may not always give benefit tothe overall capacity in the system Considering an examplethat if the traffic load in macrocell is low and there areenough radio resources that can be allocated to the servedusers the throughput of these users is consequently limited bytheir channel conditions that is SINRs In this case forcingthese users into the picocell by CRE can further damage theSINRs and consequently damage the usersrsquo throughput Onthe other hand if the traffic load in macrocell is high and theusersrsquo throughputs are limited by the available radio resourcehanding over these users to LPN by CRE can improve theusersrsquo throughputs
To get the optimal overall capacity in the next section anadaptive algorithm to optimize theCSO setting via predictionof overall capacity for different CSO values is proposed
3 Capacity Optimization Algorithm withAdaptive Cell Range Control
In this section an adaptive algorithm is proposed to optimizethe overall capacity by adjusting the CSO setting In thisalgorithm the CSO setting is updated periodically At eachupdate instance the current CSO value is updated by a newvalue if the predicted overall capacity can be increased by thisnew attempting valueThe attemptingCSO setting is obtainedby increasing or decreasing a step to the current CSO Thenew CSO value which gives the best overall capacity is thenselected The algorithm is described step by step as in below
31 Step 1 User Grouping with Attempting CSO SettingsSuppose CSO updating step is ΔCSO At each update instance119905 CSO value is updated based on current CSO value CSO
119905minus1
whichwas set at time interval 119905minus1The attemptingCSO valueCSO119905att is obtained by either increasing or decreasingCSO119905minus1
by the updating step ΔCSO as
CSO119905att = CSO119905inc = CSO119905minus1 + ΔCSO (40a)
CSO119905att = CSO119905dec = CSO119905minus1 minus ΔCSO (40b)
Based on the RSRP measurements from both macrocelland picocell for the attempted CSO setting all of the userscan be divided into three groups
Group 1Macro users These users stay in the macrocell bothwith the previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att le RSRP119898
119906
(41)
Group 2 Pico usersThese users stay in the picocell both withthe previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1gt RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(42)
Group 3 Users in CRE region These users will change theirserving cell with new CSO setting Take the case that CSO isincreased as the following example
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(43)
On the basis of the user grouping the capacities of usersin different groups are predicted respectively and the overallcapacity is predicted consequently
32 Step 2 Capacity Prediction with Attempting CSO SettingsIn step 2 the capacities of users in different groups arepredicted respectively according to the predicted SINR andcell load More details can be found below
321 Step 21 Capacity Prediction for Users in CRE RegionTo estimate the new capacity of users in CRE region we needto estimate the new SINR and predict the new cell load inpicocell and in macrocell
Taking the case that the CSO is increased by a step asexample to estimate the new SINR in picocell 120574(119901)
119906CRE we canuse
120574(119901)
119906CRE =RSRP(119901)
(RSRP(119898)120574(119898)119906CRE) + RSRP
(119898)minus RSRP(119901)
(44)
To estimate the new picocell load we need to estimate therequired number of RBs according to the new SINR For CBRservice this can be estimated similarly to (26) by
119873119906=
119863119906
119877 (120574(119901)
119906CRE) sdot BW (45)
For services other than CBR service the new requirednumber of RBs can be estimated according to the relative rela-tionship between the previous SINR in macrocell previousrequired number of RBs inmacrocell and the predicted SINRin picocell
119873119906
new= 119873119906
oldsdot
119877 (120574(119898)
119906CRE)
119877 (120574(119901)
119906CRE) (46)
Based on the prediction of the required number of RBsthe new picocell load can be estimated by assuming therequired number of RBs is unchanged for other users inpicocell
120588119901
new=
sum119873119875
119906=1119873119906+ sum119873119877
119906=1119873119906
new
NRBavail119901
(47)
8 International Journal of Antennas and Propagation
Consequently the new capacity of users newly added toCRE region can be estimated according to the analysis inSection 2
322 Step 22 Capacity Prediction for Macro Users and PicoUsers Formacro users to predict the SINR of users with newCSO the picocell load with new CSO needs to be estimatedThis can be achieved by method described in Step 21
The new cell load in macrocell can be simply estimatedby the new number of served users and assuming that therequired numbers of RBs for macro users are unchanged
The capacity of users in picocell can be estimated simi-larly
33 Optimal CSO Selection and Updating The predictednew capacities of users in different groups are summedtogether to get the overall predicted capacity The predictedoverall capacity is then compared with current capacity If theestimated new capacity is larger than the current capacity theCSO is updated
If 119862 (CSO119905att) gt 119862 (CSO119905minus1) CSO119905 = CSO119905att (48)
Otherwise CSO keeps unchangedIf the CSO is decreased by a step the overall capacity
can be estimated similarly by reversing the calculation shownabove If the new CSO value gives better predicted perfor-mance the CSO value is updated
With the proposed adaptive algorithm the CSO value isadjusted according to the predicted overall capacity in bothmacrocell and picocell By selecting the most suitable CSOvalue the overall capacity is expected to be optimized
4 Numerical Results
The proposed adaptive CSO updating algorithm is evaluatedvia system-level simulations In this section the simulationresults are presented and discussed The simulation assump-tions and parameter settings are illustrated in Table 1
Two typical scenarios are tested in the evaluation Inscenario 1 the users are uniformly distributed in the entiresimulated area LPNs are deployed randomly In scenario 2LPNs are deployed in hotspot areas and the hotspot areastake nearly 60of the usersThe base station deployment anduser distribution in two scenarios are illustrated in Figures 3and 4 To have a clear view only part of the simulated areas isillustrated
The percentages of users belonging to macrocell or LPNsin the two scenarios are illustrated in Figure 5 with differentCSO settings It can be observed that when CSO settingincreases the users belonging to LPNs increase It can beobserved as well that the percentages of users belonging toLPNs in hotspot scenario are much higher than in uniformscenario
Figure 6 shows the change of average cell load withdifferent CSO settings in different scenarios In uniformscenario most of the users are in macrocells and the loadin the macrocell is consequently quite high the cell load inpicocell increases with the increase in the CSO setting In
0
1
2
3
4
5
6
911
15 17
Figure 3 Cell layout and user distribution in uniform scenario
0
1
2
3
4
5
6
911
15 17
Figure 4 Cell layout and user distribution in hotspot scenario
hotspot scenario the average cell load inmacrocells decreaseswith the increase in CSO setting while the average cell load inLPNs increases further with the increased CSO setting
In Figure 7 the average sector throughput gain withdifferent CSO settings is compared with the proposed adap-tive algorithm for uniform scenario The average sectorthroughput with CSO = 0 dB is used as reference to obtainthe average sector throughput gain It can be observed that inthis scenario a higherCSO setting is desirable to offloadmoreusers to LPNs The proposed adaptive algorithm can achievethe similar throughput to the throughput of the optimal CSOsetting
In Figure 8 the average sector throughput gain with dif-ferent CSO settings is compared with the proposed adaptivealgorithm in hotspot scenarioThe average sector throughput
International Journal of Antennas and Propagation 9
0001020304050607080910
Uniform-macroUniform-pico
Hotspot-picoHotspot-macro
CSO 15CSO 10CSO 5CSO 0
Con
nect
ion
ratio
Figure 5 Connection ratio with different CSOs
0 2 4 6 8 10
05
06
07
08
09
10
Uniform-macroUniform-pico
Hotspot-macroHotspot-pico
CSO (dB)
Nor
mal
ized
cell
load
Figure 6 Cell load with different CSOs
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 7 Average sector throughput gain in uniform scenario
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 8 Average sector throughput gain in hotspot scenario
with CSO = 0 dB is used as reference to obtain the averagesector throughput gain In this scenario CSO of 10 dB isdesirable to balance the offloading gain and SINR loss Theproposed algorithm achieves the similar performance to theoptimal CSO setting
In different scenarios the optimal CSO setting is differ-ent However the proposed adaptive algorithm can achievesimilar performance with the optimal CSO setting in bothscenarios
5 Conclusion
As an attractivemeans of expandingmobile network capacityheterogeneous network has been included in LTE-advancedCell range extension is an important feature in heterogeneousnetwork to improve the utilization of the resource of lowpower nodes However the users in the cell range extensionarea usually suffer from high interference from macrocellsTo avoid the unnecessary interference it is desirable tocarefully consider the optimal cell selection offset value Inthis paper the capacity of users in macrocell and small cellsis analyzed thoroughly The conditions to improve capacityvia cell range extension are derived respectively for usersboth in macrocell and in small cell Based on the analysis anadaptive small cell coverage control algorithm is proposedWith the proposed adaptive algorithm the cell selectionoffset is updated periodically based on the prediction of theoverall system capacityThe algorithm is evaluated by system-level simulations and the results indicate that with theproposed algorithm a nearly optimal system performancecan be achieved in all tested load cases
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
10 International Journal of Antennas and Propagation
Acknowledgments
This work was supported by the State Major Science andTechnology Special Projects (Grant no 2013ZX03001026-001) and the Fundamental Research Funds for the CentralUniversities (Grant no 2014RC0107)
References
[1] A Damnjanovic J Montojo Y Wei et al ldquoA survey on 3GPPheterogeneous networksrdquo IEEE Wireless Communications vol18 no 3 pp 10ndash21 2011
[2] E Dahlman S Parkvall and J Skold 4G LTELTE-Advanced forMobile Broadband Academic Press 2011
[3] A Ghosh R Ratasuk B Mondal N Mangalvedhe and TThomas ldquoLTE-advanced next-generation wireless broadbandtechnologyrdquo IEEE Wireless Communications vol 17 no 3 pp10ndash22 2010
[4] X Zhang XGuW Li L Zhang J Shen andYWan ldquoThe studyof indoor and field trials on 2times8MIMOarchitecture in TD-LTEnetworkrdquo International Journal of Antennas and Propagationsvol 2013 Article ID 181579 9 pages 2013
[5] G Yuan X Zhang WWang and Y Yang ldquoCarrier aggregationfor LTE-advancedmobile communication systemsrdquo IEEE Com-munications Magazine vol 48 no 2 pp 88ndash93 2010
[6] R Irmer H Droste P Marsch et al ldquoCoordinated multipointconcepts performance and field trial resultsrdquo IEEE Communi-cations Magazine vol 49 no 2 pp 102ndash111 2011
[7] I Siomina and Y Di ldquoLoad balancing in heterogeneous LTErange optimization via cell offset and load-coupling character-izationrdquo in Proceedings of the IEEE International Conference onCommunications pp 1357ndash1361 2012
[8] The 3rd Generation Partnership Project (3GPP) ldquoFeasibilitystudy for further advancements for E-UTRA (LTEAdvanced)(Release 10)rdquo Tech Rep TR 36912 2011
[9] The 3rd Generation Partnership Project(3GPP) ldquoSystem per-formance of heterogeneous networks with range expansionrdquoTech Rep R1-101203 Samsung 2010
[10] I Guvenc ldquoCapacity and fairness analysis of heterogeneous net-works with range expansion and interference coordinationrdquoIEEECommunications Letters vol 15 no 10 pp 1084ndash1087 2011
[11] H-S Jo Y J Sang P Xia and J G Andrews ldquoHeterogeneouscellular networks with flexible cell association a comprehensivedownlink SINR analysisrdquo IEEE Transactions on Wireless Com-munications vol 11 no 10 pp 3484ndash3495 2012
[12] ldquoSelf-optimisationand self-configuration in wireless networksrdquoSOCRATES European Research Project httpwwwfp7-socratesorg
[13] The 3rd Generation Partnership Project(3GPP) ldquoSelf-config-uring and self-optimizing network use cases and solutionsrdquoTech Rep TR 36902 2009
[14] ldquoUse cases related to Self Organizing network Overall descrip-tionrdquoNext generationMobileNetworks httpwwwngmnorg
[15] X Chu and D Lopez-Perez Heterogeneous Cellular NetworksTheory Simulation and Deployment Cambridge UniversityPress 2013
[16] P Tian H Tian J Zhu L Chen and X She ldquoAn adaptive biasconfiguration strategy for range extension in LTE-Advancedheterogeneous networksrdquo inProceedings of the IET InternationalConference on Communication Technology and Application(ICCTA rsquo11) pp 336ndash340 2011
[17] K Kikuchi and H Otsuka ldquoProposal of adaptive control CREin heterogeneous networksrdquo in Proceedings of the IEEE Inter-national Symposium on Personal Indoor and Mobile RadioCommunications (PIMRC rsquo12) pp 910ndash914 2012
[18] I Viering M Dottling and A Lobinger ldquoA mathematical per-spective of self-optimizing wireless networksrdquo in Proceedings ofthe IEEE International Conference on Communications (ICCrsquo09) pp 1ndash6 June 2009
[19] A Lobinger S Stefanski T Jansen and I Balan ldquoLoad balanc-ing in downlink LTE self-optimizing networksrdquo in Proceedingsof the IEEE 71st Vehicular Technology Conference (VTC rsquo10) pp1ndash5 May 2010
[20] ldquoGuidelines for evaluation of radio transmission technologiesfor IMT-2000rdquo Recommendation ITU-R M1225 1997
[21] The 3rd Generation Partnership Project (3GPP) ldquoFurtheradvancements for E-UTRAN physical layer aspects (Release9)rdquo Tech Rep TS 36814 2010
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Antennas and Propagation 7
From the analysis above it can be observed that withCRE the overall capacity not only depends on the change ofSINRs of users in macrocell in picocell and in CRE regionbut also depends on the load situation and change in bothmacrocell and picocell CRE may not always give benefit tothe overall capacity in the system Considering an examplethat if the traffic load in macrocell is low and there areenough radio resources that can be allocated to the servedusers the throughput of these users is consequently limited bytheir channel conditions that is SINRs In this case forcingthese users into the picocell by CRE can further damage theSINRs and consequently damage the usersrsquo throughput Onthe other hand if the traffic load in macrocell is high and theusersrsquo throughputs are limited by the available radio resourcehanding over these users to LPN by CRE can improve theusersrsquo throughputs
To get the optimal overall capacity in the next section anadaptive algorithm to optimize theCSO setting via predictionof overall capacity for different CSO values is proposed
3 Capacity Optimization Algorithm withAdaptive Cell Range Control
In this section an adaptive algorithm is proposed to optimizethe overall capacity by adjusting the CSO setting In thisalgorithm the CSO setting is updated periodically At eachupdate instance the current CSO value is updated by a newvalue if the predicted overall capacity can be increased by thisnew attempting valueThe attemptingCSO setting is obtainedby increasing or decreasing a step to the current CSO Thenew CSO value which gives the best overall capacity is thenselected The algorithm is described step by step as in below
31 Step 1 User Grouping with Attempting CSO SettingsSuppose CSO updating step is ΔCSO At each update instance119905 CSO value is updated based on current CSO value CSO
119905minus1
whichwas set at time interval 119905minus1The attemptingCSO valueCSO119905att is obtained by either increasing or decreasingCSO119905minus1
by the updating step ΔCSO as
CSO119905att = CSO119905inc = CSO119905minus1 + ΔCSO (40a)
CSO119905att = CSO119905dec = CSO119905minus1 minus ΔCSO (40b)
Based on the RSRP measurements from both macrocelland picocell for the attempted CSO setting all of the userscan be divided into three groups
Group 1Macro users These users stay in the macrocell bothwith the previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att le RSRP119898
119906
(41)
Group 2 Pico usersThese users stay in the picocell both withthe previous CSO setting and the new CSO setting
RSRP119901119906+ CSO
119905minus1gt RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(42)
Group 3 Users in CRE region These users will change theirserving cell with new CSO setting Take the case that CSO isincreased as the following example
RSRP119901119906+ CSO
119905minus1le RSRP119898
119906
RSRP119901119906+ CSO
119905att gt RSRP119898
119906
(43)
On the basis of the user grouping the capacities of usersin different groups are predicted respectively and the overallcapacity is predicted consequently
32 Step 2 Capacity Prediction with Attempting CSO SettingsIn step 2 the capacities of users in different groups arepredicted respectively according to the predicted SINR andcell load More details can be found below
321 Step 21 Capacity Prediction for Users in CRE RegionTo estimate the new capacity of users in CRE region we needto estimate the new SINR and predict the new cell load inpicocell and in macrocell
Taking the case that the CSO is increased by a step asexample to estimate the new SINR in picocell 120574(119901)
119906CRE we canuse
120574(119901)
119906CRE =RSRP(119901)
(RSRP(119898)120574(119898)119906CRE) + RSRP
(119898)minus RSRP(119901)
(44)
To estimate the new picocell load we need to estimate therequired number of RBs according to the new SINR For CBRservice this can be estimated similarly to (26) by
119873119906=
119863119906
119877 (120574(119901)
119906CRE) sdot BW (45)
For services other than CBR service the new requirednumber of RBs can be estimated according to the relative rela-tionship between the previous SINR in macrocell previousrequired number of RBs inmacrocell and the predicted SINRin picocell
119873119906
new= 119873119906
oldsdot
119877 (120574(119898)
119906CRE)
119877 (120574(119901)
119906CRE) (46)
Based on the prediction of the required number of RBsthe new picocell load can be estimated by assuming therequired number of RBs is unchanged for other users inpicocell
120588119901
new=
sum119873119875
119906=1119873119906+ sum119873119877
119906=1119873119906
new
NRBavail119901
(47)
8 International Journal of Antennas and Propagation
Consequently the new capacity of users newly added toCRE region can be estimated according to the analysis inSection 2
322 Step 22 Capacity Prediction for Macro Users and PicoUsers Formacro users to predict the SINR of users with newCSO the picocell load with new CSO needs to be estimatedThis can be achieved by method described in Step 21
The new cell load in macrocell can be simply estimatedby the new number of served users and assuming that therequired numbers of RBs for macro users are unchanged
The capacity of users in picocell can be estimated simi-larly
33 Optimal CSO Selection and Updating The predictednew capacities of users in different groups are summedtogether to get the overall predicted capacity The predictedoverall capacity is then compared with current capacity If theestimated new capacity is larger than the current capacity theCSO is updated
If 119862 (CSO119905att) gt 119862 (CSO119905minus1) CSO119905 = CSO119905att (48)
Otherwise CSO keeps unchangedIf the CSO is decreased by a step the overall capacity
can be estimated similarly by reversing the calculation shownabove If the new CSO value gives better predicted perfor-mance the CSO value is updated
With the proposed adaptive algorithm the CSO value isadjusted according to the predicted overall capacity in bothmacrocell and picocell By selecting the most suitable CSOvalue the overall capacity is expected to be optimized
4 Numerical Results
The proposed adaptive CSO updating algorithm is evaluatedvia system-level simulations In this section the simulationresults are presented and discussed The simulation assump-tions and parameter settings are illustrated in Table 1
Two typical scenarios are tested in the evaluation Inscenario 1 the users are uniformly distributed in the entiresimulated area LPNs are deployed randomly In scenario 2LPNs are deployed in hotspot areas and the hotspot areastake nearly 60of the usersThe base station deployment anduser distribution in two scenarios are illustrated in Figures 3and 4 To have a clear view only part of the simulated areas isillustrated
The percentages of users belonging to macrocell or LPNsin the two scenarios are illustrated in Figure 5 with differentCSO settings It can be observed that when CSO settingincreases the users belonging to LPNs increase It can beobserved as well that the percentages of users belonging toLPNs in hotspot scenario are much higher than in uniformscenario
Figure 6 shows the change of average cell load withdifferent CSO settings in different scenarios In uniformscenario most of the users are in macrocells and the loadin the macrocell is consequently quite high the cell load inpicocell increases with the increase in the CSO setting In
0
1
2
3
4
5
6
911
15 17
Figure 3 Cell layout and user distribution in uniform scenario
0
1
2
3
4
5
6
911
15 17
Figure 4 Cell layout and user distribution in hotspot scenario
hotspot scenario the average cell load inmacrocells decreaseswith the increase in CSO setting while the average cell load inLPNs increases further with the increased CSO setting
In Figure 7 the average sector throughput gain withdifferent CSO settings is compared with the proposed adap-tive algorithm for uniform scenario The average sectorthroughput with CSO = 0 dB is used as reference to obtainthe average sector throughput gain It can be observed that inthis scenario a higherCSO setting is desirable to offloadmoreusers to LPNs The proposed adaptive algorithm can achievethe similar throughput to the throughput of the optimal CSOsetting
In Figure 8 the average sector throughput gain with dif-ferent CSO settings is compared with the proposed adaptivealgorithm in hotspot scenarioThe average sector throughput
International Journal of Antennas and Propagation 9
0001020304050607080910
Uniform-macroUniform-pico
Hotspot-picoHotspot-macro
CSO 15CSO 10CSO 5CSO 0
Con
nect
ion
ratio
Figure 5 Connection ratio with different CSOs
0 2 4 6 8 10
05
06
07
08
09
10
Uniform-macroUniform-pico
Hotspot-macroHotspot-pico
CSO (dB)
Nor
mal
ized
cell
load
Figure 6 Cell load with different CSOs
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 7 Average sector throughput gain in uniform scenario
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 8 Average sector throughput gain in hotspot scenario
with CSO = 0 dB is used as reference to obtain the averagesector throughput gain In this scenario CSO of 10 dB isdesirable to balance the offloading gain and SINR loss Theproposed algorithm achieves the similar performance to theoptimal CSO setting
In different scenarios the optimal CSO setting is differ-ent However the proposed adaptive algorithm can achievesimilar performance with the optimal CSO setting in bothscenarios
5 Conclusion
As an attractivemeans of expandingmobile network capacityheterogeneous network has been included in LTE-advancedCell range extension is an important feature in heterogeneousnetwork to improve the utilization of the resource of lowpower nodes However the users in the cell range extensionarea usually suffer from high interference from macrocellsTo avoid the unnecessary interference it is desirable tocarefully consider the optimal cell selection offset value Inthis paper the capacity of users in macrocell and small cellsis analyzed thoroughly The conditions to improve capacityvia cell range extension are derived respectively for usersboth in macrocell and in small cell Based on the analysis anadaptive small cell coverage control algorithm is proposedWith the proposed adaptive algorithm the cell selectionoffset is updated periodically based on the prediction of theoverall system capacityThe algorithm is evaluated by system-level simulations and the results indicate that with theproposed algorithm a nearly optimal system performancecan be achieved in all tested load cases
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
10 International Journal of Antennas and Propagation
Acknowledgments
This work was supported by the State Major Science andTechnology Special Projects (Grant no 2013ZX03001026-001) and the Fundamental Research Funds for the CentralUniversities (Grant no 2014RC0107)
References
[1] A Damnjanovic J Montojo Y Wei et al ldquoA survey on 3GPPheterogeneous networksrdquo IEEE Wireless Communications vol18 no 3 pp 10ndash21 2011
[2] E Dahlman S Parkvall and J Skold 4G LTELTE-Advanced forMobile Broadband Academic Press 2011
[3] A Ghosh R Ratasuk B Mondal N Mangalvedhe and TThomas ldquoLTE-advanced next-generation wireless broadbandtechnologyrdquo IEEE Wireless Communications vol 17 no 3 pp10ndash22 2010
[4] X Zhang XGuW Li L Zhang J Shen andYWan ldquoThe studyof indoor and field trials on 2times8MIMOarchitecture in TD-LTEnetworkrdquo International Journal of Antennas and Propagationsvol 2013 Article ID 181579 9 pages 2013
[5] G Yuan X Zhang WWang and Y Yang ldquoCarrier aggregationfor LTE-advancedmobile communication systemsrdquo IEEE Com-munications Magazine vol 48 no 2 pp 88ndash93 2010
[6] R Irmer H Droste P Marsch et al ldquoCoordinated multipointconcepts performance and field trial resultsrdquo IEEE Communi-cations Magazine vol 49 no 2 pp 102ndash111 2011
[7] I Siomina and Y Di ldquoLoad balancing in heterogeneous LTErange optimization via cell offset and load-coupling character-izationrdquo in Proceedings of the IEEE International Conference onCommunications pp 1357ndash1361 2012
[8] The 3rd Generation Partnership Project (3GPP) ldquoFeasibilitystudy for further advancements for E-UTRA (LTEAdvanced)(Release 10)rdquo Tech Rep TR 36912 2011
[9] The 3rd Generation Partnership Project(3GPP) ldquoSystem per-formance of heterogeneous networks with range expansionrdquoTech Rep R1-101203 Samsung 2010
[10] I Guvenc ldquoCapacity and fairness analysis of heterogeneous net-works with range expansion and interference coordinationrdquoIEEECommunications Letters vol 15 no 10 pp 1084ndash1087 2011
[11] H-S Jo Y J Sang P Xia and J G Andrews ldquoHeterogeneouscellular networks with flexible cell association a comprehensivedownlink SINR analysisrdquo IEEE Transactions on Wireless Com-munications vol 11 no 10 pp 3484ndash3495 2012
[12] ldquoSelf-optimisationand self-configuration in wireless networksrdquoSOCRATES European Research Project httpwwwfp7-socratesorg
[13] The 3rd Generation Partnership Project(3GPP) ldquoSelf-config-uring and self-optimizing network use cases and solutionsrdquoTech Rep TR 36902 2009
[14] ldquoUse cases related to Self Organizing network Overall descrip-tionrdquoNext generationMobileNetworks httpwwwngmnorg
[15] X Chu and D Lopez-Perez Heterogeneous Cellular NetworksTheory Simulation and Deployment Cambridge UniversityPress 2013
[16] P Tian H Tian J Zhu L Chen and X She ldquoAn adaptive biasconfiguration strategy for range extension in LTE-Advancedheterogeneous networksrdquo inProceedings of the IET InternationalConference on Communication Technology and Application(ICCTA rsquo11) pp 336ndash340 2011
[17] K Kikuchi and H Otsuka ldquoProposal of adaptive control CREin heterogeneous networksrdquo in Proceedings of the IEEE Inter-national Symposium on Personal Indoor and Mobile RadioCommunications (PIMRC rsquo12) pp 910ndash914 2012
[18] I Viering M Dottling and A Lobinger ldquoA mathematical per-spective of self-optimizing wireless networksrdquo in Proceedings ofthe IEEE International Conference on Communications (ICCrsquo09) pp 1ndash6 June 2009
[19] A Lobinger S Stefanski T Jansen and I Balan ldquoLoad balanc-ing in downlink LTE self-optimizing networksrdquo in Proceedingsof the IEEE 71st Vehicular Technology Conference (VTC rsquo10) pp1ndash5 May 2010
[20] ldquoGuidelines for evaluation of radio transmission technologiesfor IMT-2000rdquo Recommendation ITU-R M1225 1997
[21] The 3rd Generation Partnership Project (3GPP) ldquoFurtheradvancements for E-UTRAN physical layer aspects (Release9)rdquo Tech Rep TS 36814 2010
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
8 International Journal of Antennas and Propagation
Consequently the new capacity of users newly added toCRE region can be estimated according to the analysis inSection 2
322 Step 22 Capacity Prediction for Macro Users and PicoUsers Formacro users to predict the SINR of users with newCSO the picocell load with new CSO needs to be estimatedThis can be achieved by method described in Step 21
The new cell load in macrocell can be simply estimatedby the new number of served users and assuming that therequired numbers of RBs for macro users are unchanged
The capacity of users in picocell can be estimated simi-larly
33 Optimal CSO Selection and Updating The predictednew capacities of users in different groups are summedtogether to get the overall predicted capacity The predictedoverall capacity is then compared with current capacity If theestimated new capacity is larger than the current capacity theCSO is updated
If 119862 (CSO119905att) gt 119862 (CSO119905minus1) CSO119905 = CSO119905att (48)
Otherwise CSO keeps unchangedIf the CSO is decreased by a step the overall capacity
can be estimated similarly by reversing the calculation shownabove If the new CSO value gives better predicted perfor-mance the CSO value is updated
With the proposed adaptive algorithm the CSO value isadjusted according to the predicted overall capacity in bothmacrocell and picocell By selecting the most suitable CSOvalue the overall capacity is expected to be optimized
4 Numerical Results
The proposed adaptive CSO updating algorithm is evaluatedvia system-level simulations In this section the simulationresults are presented and discussed The simulation assump-tions and parameter settings are illustrated in Table 1
Two typical scenarios are tested in the evaluation Inscenario 1 the users are uniformly distributed in the entiresimulated area LPNs are deployed randomly In scenario 2LPNs are deployed in hotspot areas and the hotspot areastake nearly 60of the usersThe base station deployment anduser distribution in two scenarios are illustrated in Figures 3and 4 To have a clear view only part of the simulated areas isillustrated
The percentages of users belonging to macrocell or LPNsin the two scenarios are illustrated in Figure 5 with differentCSO settings It can be observed that when CSO settingincreases the users belonging to LPNs increase It can beobserved as well that the percentages of users belonging toLPNs in hotspot scenario are much higher than in uniformscenario
Figure 6 shows the change of average cell load withdifferent CSO settings in different scenarios In uniformscenario most of the users are in macrocells and the loadin the macrocell is consequently quite high the cell load inpicocell increases with the increase in the CSO setting In
0
1
2
3
4
5
6
911
15 17
Figure 3 Cell layout and user distribution in uniform scenario
0
1
2
3
4
5
6
911
15 17
Figure 4 Cell layout and user distribution in hotspot scenario
hotspot scenario the average cell load inmacrocells decreaseswith the increase in CSO setting while the average cell load inLPNs increases further with the increased CSO setting
In Figure 7 the average sector throughput gain withdifferent CSO settings is compared with the proposed adap-tive algorithm for uniform scenario The average sectorthroughput with CSO = 0 dB is used as reference to obtainthe average sector throughput gain It can be observed that inthis scenario a higherCSO setting is desirable to offloadmoreusers to LPNs The proposed adaptive algorithm can achievethe similar throughput to the throughput of the optimal CSOsetting
In Figure 8 the average sector throughput gain with dif-ferent CSO settings is compared with the proposed adaptivealgorithm in hotspot scenarioThe average sector throughput
International Journal of Antennas and Propagation 9
0001020304050607080910
Uniform-macroUniform-pico
Hotspot-picoHotspot-macro
CSO 15CSO 10CSO 5CSO 0
Con
nect
ion
ratio
Figure 5 Connection ratio with different CSOs
0 2 4 6 8 10
05
06
07
08
09
10
Uniform-macroUniform-pico
Hotspot-macroHotspot-pico
CSO (dB)
Nor
mal
ized
cell
load
Figure 6 Cell load with different CSOs
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 7 Average sector throughput gain in uniform scenario
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 8 Average sector throughput gain in hotspot scenario
with CSO = 0 dB is used as reference to obtain the averagesector throughput gain In this scenario CSO of 10 dB isdesirable to balance the offloading gain and SINR loss Theproposed algorithm achieves the similar performance to theoptimal CSO setting
In different scenarios the optimal CSO setting is differ-ent However the proposed adaptive algorithm can achievesimilar performance with the optimal CSO setting in bothscenarios
5 Conclusion
As an attractivemeans of expandingmobile network capacityheterogeneous network has been included in LTE-advancedCell range extension is an important feature in heterogeneousnetwork to improve the utilization of the resource of lowpower nodes However the users in the cell range extensionarea usually suffer from high interference from macrocellsTo avoid the unnecessary interference it is desirable tocarefully consider the optimal cell selection offset value Inthis paper the capacity of users in macrocell and small cellsis analyzed thoroughly The conditions to improve capacityvia cell range extension are derived respectively for usersboth in macrocell and in small cell Based on the analysis anadaptive small cell coverage control algorithm is proposedWith the proposed adaptive algorithm the cell selectionoffset is updated periodically based on the prediction of theoverall system capacityThe algorithm is evaluated by system-level simulations and the results indicate that with theproposed algorithm a nearly optimal system performancecan be achieved in all tested load cases
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
10 International Journal of Antennas and Propagation
Acknowledgments
This work was supported by the State Major Science andTechnology Special Projects (Grant no 2013ZX03001026-001) and the Fundamental Research Funds for the CentralUniversities (Grant no 2014RC0107)
References
[1] A Damnjanovic J Montojo Y Wei et al ldquoA survey on 3GPPheterogeneous networksrdquo IEEE Wireless Communications vol18 no 3 pp 10ndash21 2011
[2] E Dahlman S Parkvall and J Skold 4G LTELTE-Advanced forMobile Broadband Academic Press 2011
[3] A Ghosh R Ratasuk B Mondal N Mangalvedhe and TThomas ldquoLTE-advanced next-generation wireless broadbandtechnologyrdquo IEEE Wireless Communications vol 17 no 3 pp10ndash22 2010
[4] X Zhang XGuW Li L Zhang J Shen andYWan ldquoThe studyof indoor and field trials on 2times8MIMOarchitecture in TD-LTEnetworkrdquo International Journal of Antennas and Propagationsvol 2013 Article ID 181579 9 pages 2013
[5] G Yuan X Zhang WWang and Y Yang ldquoCarrier aggregationfor LTE-advancedmobile communication systemsrdquo IEEE Com-munications Magazine vol 48 no 2 pp 88ndash93 2010
[6] R Irmer H Droste P Marsch et al ldquoCoordinated multipointconcepts performance and field trial resultsrdquo IEEE Communi-cations Magazine vol 49 no 2 pp 102ndash111 2011
[7] I Siomina and Y Di ldquoLoad balancing in heterogeneous LTErange optimization via cell offset and load-coupling character-izationrdquo in Proceedings of the IEEE International Conference onCommunications pp 1357ndash1361 2012
[8] The 3rd Generation Partnership Project (3GPP) ldquoFeasibilitystudy for further advancements for E-UTRA (LTEAdvanced)(Release 10)rdquo Tech Rep TR 36912 2011
[9] The 3rd Generation Partnership Project(3GPP) ldquoSystem per-formance of heterogeneous networks with range expansionrdquoTech Rep R1-101203 Samsung 2010
[10] I Guvenc ldquoCapacity and fairness analysis of heterogeneous net-works with range expansion and interference coordinationrdquoIEEECommunications Letters vol 15 no 10 pp 1084ndash1087 2011
[11] H-S Jo Y J Sang P Xia and J G Andrews ldquoHeterogeneouscellular networks with flexible cell association a comprehensivedownlink SINR analysisrdquo IEEE Transactions on Wireless Com-munications vol 11 no 10 pp 3484ndash3495 2012
[12] ldquoSelf-optimisationand self-configuration in wireless networksrdquoSOCRATES European Research Project httpwwwfp7-socratesorg
[13] The 3rd Generation Partnership Project(3GPP) ldquoSelf-config-uring and self-optimizing network use cases and solutionsrdquoTech Rep TR 36902 2009
[14] ldquoUse cases related to Self Organizing network Overall descrip-tionrdquoNext generationMobileNetworks httpwwwngmnorg
[15] X Chu and D Lopez-Perez Heterogeneous Cellular NetworksTheory Simulation and Deployment Cambridge UniversityPress 2013
[16] P Tian H Tian J Zhu L Chen and X She ldquoAn adaptive biasconfiguration strategy for range extension in LTE-Advancedheterogeneous networksrdquo inProceedings of the IET InternationalConference on Communication Technology and Application(ICCTA rsquo11) pp 336ndash340 2011
[17] K Kikuchi and H Otsuka ldquoProposal of adaptive control CREin heterogeneous networksrdquo in Proceedings of the IEEE Inter-national Symposium on Personal Indoor and Mobile RadioCommunications (PIMRC rsquo12) pp 910ndash914 2012
[18] I Viering M Dottling and A Lobinger ldquoA mathematical per-spective of self-optimizing wireless networksrdquo in Proceedings ofthe IEEE International Conference on Communications (ICCrsquo09) pp 1ndash6 June 2009
[19] A Lobinger S Stefanski T Jansen and I Balan ldquoLoad balanc-ing in downlink LTE self-optimizing networksrdquo in Proceedingsof the IEEE 71st Vehicular Technology Conference (VTC rsquo10) pp1ndash5 May 2010
[20] ldquoGuidelines for evaluation of radio transmission technologiesfor IMT-2000rdquo Recommendation ITU-R M1225 1997
[21] The 3rd Generation Partnership Project (3GPP) ldquoFurtheradvancements for E-UTRAN physical layer aspects (Release9)rdquo Tech Rep TS 36814 2010
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Antennas and Propagation 9
0001020304050607080910
Uniform-macroUniform-pico
Hotspot-picoHotspot-macro
CSO 15CSO 10CSO 5CSO 0
Con
nect
ion
ratio
Figure 5 Connection ratio with different CSOs
0 2 4 6 8 10
05
06
07
08
09
10
Uniform-macroUniform-pico
Hotspot-macroHotspot-pico
CSO (dB)
Nor
mal
ized
cell
load
Figure 6 Cell load with different CSOs
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 7 Average sector throughput gain in uniform scenario
0002040608101214161820
CSO CSO 15CSO 10CSO 5CSO 0
Aver
age s
ecto
r thr
ough
put g
ain
adaptive
Figure 8 Average sector throughput gain in hotspot scenario
with CSO = 0 dB is used as reference to obtain the averagesector throughput gain In this scenario CSO of 10 dB isdesirable to balance the offloading gain and SINR loss Theproposed algorithm achieves the similar performance to theoptimal CSO setting
In different scenarios the optimal CSO setting is differ-ent However the proposed adaptive algorithm can achievesimilar performance with the optimal CSO setting in bothscenarios
5 Conclusion
As an attractivemeans of expandingmobile network capacityheterogeneous network has been included in LTE-advancedCell range extension is an important feature in heterogeneousnetwork to improve the utilization of the resource of lowpower nodes However the users in the cell range extensionarea usually suffer from high interference from macrocellsTo avoid the unnecessary interference it is desirable tocarefully consider the optimal cell selection offset value Inthis paper the capacity of users in macrocell and small cellsis analyzed thoroughly The conditions to improve capacityvia cell range extension are derived respectively for usersboth in macrocell and in small cell Based on the analysis anadaptive small cell coverage control algorithm is proposedWith the proposed adaptive algorithm the cell selectionoffset is updated periodically based on the prediction of theoverall system capacityThe algorithm is evaluated by system-level simulations and the results indicate that with theproposed algorithm a nearly optimal system performancecan be achieved in all tested load cases
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
10 International Journal of Antennas and Propagation
Acknowledgments
This work was supported by the State Major Science andTechnology Special Projects (Grant no 2013ZX03001026-001) and the Fundamental Research Funds for the CentralUniversities (Grant no 2014RC0107)
References
[1] A Damnjanovic J Montojo Y Wei et al ldquoA survey on 3GPPheterogeneous networksrdquo IEEE Wireless Communications vol18 no 3 pp 10ndash21 2011
[2] E Dahlman S Parkvall and J Skold 4G LTELTE-Advanced forMobile Broadband Academic Press 2011
[3] A Ghosh R Ratasuk B Mondal N Mangalvedhe and TThomas ldquoLTE-advanced next-generation wireless broadbandtechnologyrdquo IEEE Wireless Communications vol 17 no 3 pp10ndash22 2010
[4] X Zhang XGuW Li L Zhang J Shen andYWan ldquoThe studyof indoor and field trials on 2times8MIMOarchitecture in TD-LTEnetworkrdquo International Journal of Antennas and Propagationsvol 2013 Article ID 181579 9 pages 2013
[5] G Yuan X Zhang WWang and Y Yang ldquoCarrier aggregationfor LTE-advancedmobile communication systemsrdquo IEEE Com-munications Magazine vol 48 no 2 pp 88ndash93 2010
[6] R Irmer H Droste P Marsch et al ldquoCoordinated multipointconcepts performance and field trial resultsrdquo IEEE Communi-cations Magazine vol 49 no 2 pp 102ndash111 2011
[7] I Siomina and Y Di ldquoLoad balancing in heterogeneous LTErange optimization via cell offset and load-coupling character-izationrdquo in Proceedings of the IEEE International Conference onCommunications pp 1357ndash1361 2012
[8] The 3rd Generation Partnership Project (3GPP) ldquoFeasibilitystudy for further advancements for E-UTRA (LTEAdvanced)(Release 10)rdquo Tech Rep TR 36912 2011
[9] The 3rd Generation Partnership Project(3GPP) ldquoSystem per-formance of heterogeneous networks with range expansionrdquoTech Rep R1-101203 Samsung 2010
[10] I Guvenc ldquoCapacity and fairness analysis of heterogeneous net-works with range expansion and interference coordinationrdquoIEEECommunications Letters vol 15 no 10 pp 1084ndash1087 2011
[11] H-S Jo Y J Sang P Xia and J G Andrews ldquoHeterogeneouscellular networks with flexible cell association a comprehensivedownlink SINR analysisrdquo IEEE Transactions on Wireless Com-munications vol 11 no 10 pp 3484ndash3495 2012
[12] ldquoSelf-optimisationand self-configuration in wireless networksrdquoSOCRATES European Research Project httpwwwfp7-socratesorg
[13] The 3rd Generation Partnership Project(3GPP) ldquoSelf-config-uring and self-optimizing network use cases and solutionsrdquoTech Rep TR 36902 2009
[14] ldquoUse cases related to Self Organizing network Overall descrip-tionrdquoNext generationMobileNetworks httpwwwngmnorg
[15] X Chu and D Lopez-Perez Heterogeneous Cellular NetworksTheory Simulation and Deployment Cambridge UniversityPress 2013
[16] P Tian H Tian J Zhu L Chen and X She ldquoAn adaptive biasconfiguration strategy for range extension in LTE-Advancedheterogeneous networksrdquo inProceedings of the IET InternationalConference on Communication Technology and Application(ICCTA rsquo11) pp 336ndash340 2011
[17] K Kikuchi and H Otsuka ldquoProposal of adaptive control CREin heterogeneous networksrdquo in Proceedings of the IEEE Inter-national Symposium on Personal Indoor and Mobile RadioCommunications (PIMRC rsquo12) pp 910ndash914 2012
[18] I Viering M Dottling and A Lobinger ldquoA mathematical per-spective of self-optimizing wireless networksrdquo in Proceedings ofthe IEEE International Conference on Communications (ICCrsquo09) pp 1ndash6 June 2009
[19] A Lobinger S Stefanski T Jansen and I Balan ldquoLoad balanc-ing in downlink LTE self-optimizing networksrdquo in Proceedingsof the IEEE 71st Vehicular Technology Conference (VTC rsquo10) pp1ndash5 May 2010
[20] ldquoGuidelines for evaluation of radio transmission technologiesfor IMT-2000rdquo Recommendation ITU-R M1225 1997
[21] The 3rd Generation Partnership Project (3GPP) ldquoFurtheradvancements for E-UTRAN physical layer aspects (Release9)rdquo Tech Rep TS 36814 2010
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
10 International Journal of Antennas and Propagation
Acknowledgments
This work was supported by the State Major Science andTechnology Special Projects (Grant no 2013ZX03001026-001) and the Fundamental Research Funds for the CentralUniversities (Grant no 2014RC0107)
References
[1] A Damnjanovic J Montojo Y Wei et al ldquoA survey on 3GPPheterogeneous networksrdquo IEEE Wireless Communications vol18 no 3 pp 10ndash21 2011
[2] E Dahlman S Parkvall and J Skold 4G LTELTE-Advanced forMobile Broadband Academic Press 2011
[3] A Ghosh R Ratasuk B Mondal N Mangalvedhe and TThomas ldquoLTE-advanced next-generation wireless broadbandtechnologyrdquo IEEE Wireless Communications vol 17 no 3 pp10ndash22 2010
[4] X Zhang XGuW Li L Zhang J Shen andYWan ldquoThe studyof indoor and field trials on 2times8MIMOarchitecture in TD-LTEnetworkrdquo International Journal of Antennas and Propagationsvol 2013 Article ID 181579 9 pages 2013
[5] G Yuan X Zhang WWang and Y Yang ldquoCarrier aggregationfor LTE-advancedmobile communication systemsrdquo IEEE Com-munications Magazine vol 48 no 2 pp 88ndash93 2010
[6] R Irmer H Droste P Marsch et al ldquoCoordinated multipointconcepts performance and field trial resultsrdquo IEEE Communi-cations Magazine vol 49 no 2 pp 102ndash111 2011
[7] I Siomina and Y Di ldquoLoad balancing in heterogeneous LTErange optimization via cell offset and load-coupling character-izationrdquo in Proceedings of the IEEE International Conference onCommunications pp 1357ndash1361 2012
[8] The 3rd Generation Partnership Project (3GPP) ldquoFeasibilitystudy for further advancements for E-UTRA (LTEAdvanced)(Release 10)rdquo Tech Rep TR 36912 2011
[9] The 3rd Generation Partnership Project(3GPP) ldquoSystem per-formance of heterogeneous networks with range expansionrdquoTech Rep R1-101203 Samsung 2010
[10] I Guvenc ldquoCapacity and fairness analysis of heterogeneous net-works with range expansion and interference coordinationrdquoIEEECommunications Letters vol 15 no 10 pp 1084ndash1087 2011
[11] H-S Jo Y J Sang P Xia and J G Andrews ldquoHeterogeneouscellular networks with flexible cell association a comprehensivedownlink SINR analysisrdquo IEEE Transactions on Wireless Com-munications vol 11 no 10 pp 3484ndash3495 2012
[12] ldquoSelf-optimisationand self-configuration in wireless networksrdquoSOCRATES European Research Project httpwwwfp7-socratesorg
[13] The 3rd Generation Partnership Project(3GPP) ldquoSelf-config-uring and self-optimizing network use cases and solutionsrdquoTech Rep TR 36902 2009
[14] ldquoUse cases related to Self Organizing network Overall descrip-tionrdquoNext generationMobileNetworks httpwwwngmnorg
[15] X Chu and D Lopez-Perez Heterogeneous Cellular NetworksTheory Simulation and Deployment Cambridge UniversityPress 2013
[16] P Tian H Tian J Zhu L Chen and X She ldquoAn adaptive biasconfiguration strategy for range extension in LTE-Advancedheterogeneous networksrdquo inProceedings of the IET InternationalConference on Communication Technology and Application(ICCTA rsquo11) pp 336ndash340 2011
[17] K Kikuchi and H Otsuka ldquoProposal of adaptive control CREin heterogeneous networksrdquo in Proceedings of the IEEE Inter-national Symposium on Personal Indoor and Mobile RadioCommunications (PIMRC rsquo12) pp 910ndash914 2012
[18] I Viering M Dottling and A Lobinger ldquoA mathematical per-spective of self-optimizing wireless networksrdquo in Proceedings ofthe IEEE International Conference on Communications (ICCrsquo09) pp 1ndash6 June 2009
[19] A Lobinger S Stefanski T Jansen and I Balan ldquoLoad balanc-ing in downlink LTE self-optimizing networksrdquo in Proceedingsof the IEEE 71st Vehicular Technology Conference (VTC rsquo10) pp1ndash5 May 2010
[20] ldquoGuidelines for evaluation of radio transmission technologiesfor IMT-2000rdquo Recommendation ITU-R M1225 1997
[21] The 3rd Generation Partnership Project (3GPP) ldquoFurtheradvancements for E-UTRAN physical layer aspects (Release9)rdquo Tech Rep TS 36814 2010
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of