Impact of Packet Size in Adaptive Cognitive Radio Sensor ...

Research ArticleImpact of Packet Size in Adaptive CognitiveRadio Sensor Network

Mohammed Al-Medhwahi Fazirulhisyam Hashim Borhanuddin Mohd Ali and A Sali

Department of Computer and Communication Systems Engineering amp Research Centre of Excellence for Wireless andPhotonic Networks (WiPNET) Faculty of Engineering Universiti Putra Malaysia Selangor Malaysia

Correspondence should be addressed to Mohammed Al-Medhwahi med100013yahoocom

Received 12 April 2018 Revised 18 August 2018 Accepted 27 November 2018 Published 9 December 2018

Guest Editor Jiafu Wan

Copyright copy 2018 Mohammed Al-Medhwahi et al This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

A cognitive radio sensor network (CRSN) is a solution that enables sensor nodes to opportunistically access licensed radio channelsData transmitted over a network are divided into packets Inmachine-to-machinecommunicationwhich is a heterogeneousnatureof wireless networks small-size packets are the common form of traffic Due to the nature of CRSNs small data packets will notallow a balance between optimal performance of the network and fulfilling the secondary network obligations towards the primarynetwork in terms of interference Either interference or channelrsquos underutilization would result from employing data packets ofinadequate size In this paper the appropriate packet size for adaptive CRSN is investigated by examining the performances ofsmall medium and large packet size In contrast to the trends of exploiting small packets of sizes up to 128 bytes this studydemonstrates that medium-size packets are more appropriate to yield the best performance in CRSNs Simulation results show thatpackets of size 375 bytes outperform smaller and larger packets in many CRSN protocolsThe induced delay that is partially causedby interference is decreased at the same time the channels are efficiently utilized

1 Introduction

Wireless sensor networks (WSN) applications are imple-mented in traditional and emerging applications such assecurity automated industry and e-Health Currently theimplementation of WSN applications in many new surveil-lance and monitoring services is facing a disconcertingchallenge due to the radio spectrum scarcity of free licensedbands WiFi Bluetooth cordless phones and microwaveovens technologies utilize the same radio spectrum Thiscauses a serious spectrum congestion and disruption ofa wireless network considering the close proximity of theutilized frequencies [1 2] Interference caused by the hostileradio environment results in high rates of data loss whichconsumes excessive energy and shortens the WSN networkrsquoslifetime [3] On the contrary licensed bands are beingunderutilized by licensed users which can be tapped bysecondary network users (SUs) [4]

Cognitive radio (CR) technology enables SUs to oppor-tunistically utilise the licensed channels assigned to primary

network users (PUs) Such technology aims at improvingthe spectrum utilisation and mitigating the effects of thelicense-free spectrum overcrowding [5] In cognitive radionetworks (CRNs) SUs are strictly obliged to avoid inter-ference with PUs Once a PU commences transmission onthe said channel SUs must instantly evacuate the channelknown as handoff CR uses spectrum sensing (SS) to allowSUs to determine an idle radio channel to commence datatransmission It can also defer a data transmission if thechannel of interest is busy Among many signal detectiontechniques the energy detection (ED) is the most common[6 7] In the ED technique the radio channel is consideredbusy if the energy of the detected signal exceeds a predefinedthreshold value Let 119864(119899) denote the energy of the sampledsignal received by the SU receiver 1198670 denotes the hypothesisof the absence of the PU signal while1198671 denotes its presenceThe energy of the received signal can be expressed as

119864 (119899) = 119882(119899) 1198670119882(119899) + 119878 (119899) 1198671 (1)

HindawiWireless Communications and Mobile ComputingVolume 2018 Article ID 3051204 9 pageshttpsdoiorg10115520183051204

2 Wireless Communications and Mobile Computing

PU Base Station

CRSN Base Station

PU User

Sensor

SU User

Figure 1 Typical cognitive radio sensor network

where 119882(119899) represents the additive white Gaussian noise(AWGN) and 119878(119899) represents the transmitted signal multi-plied by the channel gain

The probability that the PU channel is busy is given by

119875119900119899 = 119879119900119899119879119900119899 + 119879119900119891119891 (2)

with 119879119900119899 as the channel busy time and 119879119900119891119891 as channel idletime The probability that the channel is free 119875119900119891119891 is given as

119875119900119891119891 = 119879119900119891119891119879119900119899 + 119879119900119891119891 (3)

Probability that the channel is detected as busy when itis actually busy is called the detection probability 119875119889119890 andprobability that the channel is detected as busy when it is idleis called the false alarm probability 119875119891119886 Optimal detectioncan be achieved only if the noise power is known to the SUsThe channel status is determined based on the value of thereceived signalrsquos energy119864 and a predefined threshold value119879119903 The relationship between 119875119889119890 and 119875119891119886 with 119879119903 and 119864 is givenby (4) and (5) respectively

119875119889119890 = 119875119903 119863 = 1 | 1198671 = 119875119903 119864 gt 119879119903 | 1198671 (4)

and

119875119891119886 = 119875119903 119863 = 1 | 1198670 = 119875119903 119864 gt 119879119903 | 1198670 (5)

Emerging technology that enables the opportunisticaccess for WSNrsquos units is called the cognitive radio sensornetwork (CRSN) whose typical layout is shown in Figure 1Appropriate data packet size helps limit the interferencebetween the signals of the WSN units equipped with CRcapabilities and PU signals [8ndash10] Thus size of data packetin CRSN has a twofold importance to achieve a satisfactoryperformance and to mitigate harmful interference with PUsignal Small data packets offer maximum reliability andminimum latency for most types of data traffic especially incritical traffic Together with the growth of sensor network ininternet of things (IoT) the result is that small data packetsof sizes up to 128 bytes (1024 bits) have become the populartrend in data transmission [11 12] Subsequently overheaddata and the inefficient utilisation of resources induced byusing small-size packets in emerging heterogeneous networksare often overlooked Other substantial factors such as small-size packets have low signal noise ratios (SNRs) compared tolarge size packets in effect of noise such as thermal noise [13]and the fact that large-size packets are able to achieve higherefficiency of bandwidth utilisation [14 15] is neglected

The contributions of this study can be summarised asfollows

(i) It investigates the impact of packet size on the per-formance of CRSNs in terms of two main metricsnamely the average delay and the throughput

Wireless Communications and Mobile Computing 3

(ii) It examines the interactivity between the packet sizeand the main parameters of the radio network andshows the resultant effects on the system performance

(iii) It suggests adopting medium size by proving theoutperformance of medium-size packets compared tothe small and large packets

This study is an extension of the work in [16] and it isorganised as follows Section 2 presents the related worksSection 3 introduces the systemmodel and shows the adoptedMAC protocol Section 4 presents the mathematical modelSection 5 shows the evaluation and the results are discussedand Section 6 concludes the work

2 Related Works

Several studies have introduced solutions to improve theperformances ofMACprotocols inCRSNnetworks and somehave analysed or modelled the performance metrics of anetwork but little attention was given to the influence ofpacket size [18ndash20] Data collisions can occur during datatransmission During the data transmission phase the PU isactive until the transmission endsThe longer the SUrsquos packetduration the more collisions with the PUrsquos data that resultin decreasing the throughput of the SU Furthermore datacollision can happen between SU and existing PUs and alsobetween the SUs themselves

An analytical model was presented in [21] to determinea suitable packet size subject to CRN parameters such asnetwork traffic sensing accuracy and PUrsquos density A newformula was introduced to the normalised throughput of SUunder perfect and imperfect scenarios of the SS functionThe required complexity is not acceptable in CRSNs Aimedat maximizing the goodput another study based on carrierspectrummultiple access (CSMA)mechanism [22] proposedan analytical model to determine the optimal packet sizebased on the packet error probability the collisions betweenthe SUs themselves and the collisions between SUs and PUsThe outcomes of the study do not treat the plain structureof the end sensor nodes A similar study [23] investigatedthe impact of packet size variety on the throughput of aCSMA-based cognitive WLAN In [24] a framework wasproposed and modelled aimed at maximising the achievablethroughput while the framework in [25] was more concernedwith increasing the network lifetime The performance ofthe dynamic open spectrum sharing (DOSS) MAC protocolfor a CSMA-based CRSN was analysed and modelled in[17] which incorporates multiple channel access The studypartially investigated the effect of using long and small sizesof data frames on the performance of ad hoc-based CRSNsSuch study does not treat cluster-based networks of CRSNsthat are more suitable for heterogeneous networks

3 System Model

Network nodes are considered to be uniformly distributed in2D square area of size 119897 x 119897 The primary network includes119872 data channels (1198621 119862119872) with equal bandwidths Theoperating area is divided into clusters each consisting of

one master node and several ideal sensing nodes that arespatially correlated The ideal sensing node is self-poweredand it is considered to be heterogeneous that is producingseveral types of data packets according to how critical theobtained measurements are The traffic is classified intotwo categories namely real time (RT) traffic and non-realtime (NRT) traffic RT traffic represents the most importanttraffic containing critical data which must acquire priorityin transmission and avoid interference The master node isresponsible for collecting traffic of a group of ideal nodesand performing several main CR functions The buffer ofthe master node is assumed to be infinite and there areno limits on the number of packets it may contain Theenvironment of the licensed channels is homogeneous interms of the bandwidth and the radio physical conditionsWhile the communication inside each cluster is a single-hopvia the licensed-free spectrum the communication betweenthe master node and the base station (BS) is opportunisticvia the licensed spectrum ie through the radio channelsof the primary network A common control (CC) channelis used for the communications of the control messagesThe channels between the ideal node and the master nodeare assumed to be perfect because of the short distancesbetween them The adaptivity of the adopted network refersto the framework mechanism in which the end nodes areable to resort to the alternative workflow which is CR-basedwhen the usual WSN workflow suffers from a significantdegradation in its performance [26] In the alternative flowthe sensor node reroutes its data to the master node ratherthan routing it to its original cluster head (CH)

31 Medium Access Control Existing WSN MAC protocolssuch as IEEE 802154WirelessHART and ISA-10011a are notsuitable for CRSN networks since they lack the efficiency interms of the network heterogeneity Likewise using CRNrsquosproposed algorithms such as IEEE 80222 will result inproducing huge volume of overhead data and consumingenormous amount of sensorsrsquo precious power [27ndash29] Pli-able Cognitive MAC (PCMAC) algorithm proposed andmodelled in [16] concerns the communication betweenthe master node and the BS and aims at maintaining theideal sensorrsquos capabilities in the same time of exploitingthe CR technique The master nodes use carrier spectrummultiple access with collision avoidance (CSMACA) schemeto access the CC channel that they use to exchange thecontrol messages with the BS through Request to sendclearto send (RTSCTS) mechanism is used by the master nodeand the BS through the CC channel to negotiate the datachannel reservation This is also helpful in mitigating theeffect of the hidden node problem Inside each masternodersquos buffer packets are virtually lined up in to two virtualqueues according to their category RT and NRT First comefirst served (FCFS) discipline is adopted for scheduling thepackets in each queue The RT packets have higher priorityto be submitted earlier than the NRT packets since the latteris more delay-tolerant In e-Health applications for instanceRT traffic might represent the monitoring data of patientrsquosactivities such as the measurements of the heart and the brainwhile NRT traffic represents the ordinary traffic such as theenvironmental measurements of the room


Channel

SU 1

SU 2

Packet in the buffer

Submitted Packet

Failed Packet

Spectrum Sensing

PU (On)PU (Off)

Figure 2 Packets transmission

When the BS receives an RTS message it responds witha CTS message and assigns an unallocated channel to therequesting node The BS tunes its transceiver to the channelit has assigned to receive data Concurrently the master nodereceives the channel assignmentmessage enclosedwithin theCTS message then it tunes its transceiver into the selectedradio channel and begins to sense it until it becomes idle iespectrum hole At that time the master node can commencesubmitting its data in a burst-based mannerThemaster nodewill continue to transmit its data until either its data bufferbecomes clear or it receives no Acknowledgement (ACK)message from the BS for the latest packet

Figure 2 illustrates themechanism of packet transmissionwhere two SUs exploit the channelrsquos idle time to submit theirdata Packets are transmitted in bursts The node tries tosqueeze through as many packets as it can before the PUbecomes active again As long as the transmitting node doesnot receive its ACK message a packet may interfere with thePU signal as the second packet of the SU2 Failed packethas to be transmitted again in the next transmission cycleIn such a case the master node will continue to sense thesame channel for when the PU evacuates the channel toallow for the remaining data packets to be transmitted Sincethe master node includes its buffer size information in itssubmission request the BSwill not reassign that channel untilthe allocated node finishes its transmission The assignmentis limited in time and if the node fails to submit its data for thepredefined period of time the BS can reallocate the channelto another requesting node In such cases the former nodehas to begin a new transmission cycle starting with sendingan RTS message via the CC channel

4 Mathematical Model

The MG1 model is a model where the arrival of packets isMarkovian which occurs according to a Poisson process at

rate 120582 with a general distribution 119866 of the service time 119879through the CC channel denoted by 1 The adopted modelin this study is an improved MG1 with a non-preemptivepriority scheme inasmuch as it estimates the delay not onlyin the CC channel but also in the data channels

The service time 119879 of each packet comprises the waitingtime in the queue 119882 and the transmission time 119879119905119903 Thelatter includes the contention time in the CC channel 119879119888the licensed channelrsquos sensing time 119879119904119904 and the packetsubmission time 119879119901

119879119905119903 = 119879119888 + 119879119904119904 + 119879119901 (6)

where 119879119888 = 119863119868119865119878 + 119879119887 + 119879119903119905119904 + 119878119868119865119878 + 119879119888119905119904 DIFS and SIFSare the interframe space and short interframe space durationsof the distributed coordination function (DCF) respectively119879119903119905119904 and 119879119888119905119904 represent durations of the RTS and CTS frameswhile 119879119887 represents the backoff estimated time that can beexpressed as [30]

119879119887 = 119862119882119898119894119899 1 minus 119901 minus 2119873119901119873+12 minus 4119901 minus 12 (7)

119862119882119898119894119899 is the minimum contention window size 119901 representsthe conditional collision probability and 119873 is the number ofthe contender nodes

Since the node can arrive at the assigned channel at anymoment within the [119879119900119899 + 119879119900119891119891] period the time required toobserve the channel until it becomes idle can be estimated as

119879119904119904 = 1198791199001198992 119875119900119899 (8)

where 119875119900119899 is the licensed channelrsquos occupancy probability (byPU) obtained by (2)

Assuming that both RT and NRT packets have the samesize packet submission time can be expressed as

119879119901 = 119862119862119860 + 119871119863119889119903 + 119878119868119865119878 + 119879119886119888119896 (9)


where

119862119862119860 is clear channel assessment average time119871 is packet size119863119889119903 is data channelrsquos capacity (bitss)

For simplification the perfect sensing is assumed mean-ing that119875119889119890 = 1 and119875119891119886 = 0 However in the case of imperfectsensing the new value of the probability that a channel willbe sensed idle will be given by 119875119900119891119891(1 minus 119875119891119886) whereas the oldvalue ie for the perfect sensing 119875119900119891119891 can be obtained by (3)In several applications the acceptable values for 119875119891119886 and 119875119889119890are 01 and 09 respectively

The occupancy ratio of the CC channel 120601119888119888 is the sum ofthe occupancy ratios of RT and NRT 120601119888119888 = 120601119903119905119888 +120601119899119903119905119888 where

120601119903119905119888 = 120582119903119905119879119888and 120601119899119903119905119888 = 120582119899119903119905119888119879119888 (10)

To maintain the system stability the occupation rate for theCC channel must be less than one Similarly the occupationrate of the data channels after considering the number ofutilised channels can be estimated for RT and NRT packetsrespectively as

120601119903119905119889 = 119873120582119903119905119879119904119904 + 119879119901119862and 120601119899119903119905119889 = 119873120582119899119903119905119879119904119904 + 119879119901119862

(11)

Themeanwaiting time for anRTpacket119882119903119905 is estimatedas

119882119903119905 = 119871119903119905119879 + ΦR (12)

where R is the mean residual service time of a currentserved packet even if it is NRT and Φ represents the generaloccupation rate for the entire traffic in the CC and the datachannels The model is nonpreemptive thus

R = 11987922119879 = 1205902119879 + 1198792

2119879 (13)

By using Littlersquos Law and when 119871119903119905 represents the meansize of the RT packet queue and equals 120582119903119905119882119903119905 (12) can berewritten as

119882119903119905 (1 minus 120601119903119905) = 120601119903119905R (14)

where 120601119903119905 represents the total occupation rate for RT packetsThus the mean waiting time for an RT packet is estimatedas

119882119903119905 = ΦR1 minus 120601119903119905 (15)

and the mean throughput time ie the average delay time is

119879119903119905 = 119882119903119905 + 119879 (16)

Considering that a new packet may arrive at any momentduring the service time the mean residual service time canbe estimated as

R = 12 (1198882119904 + 1) 119879 (17)

where 119888119904 is the coefficient of the service variation time and isequal to 120590119904119879 Note that in case of exponential service time1198882119904 is equal to 1 and in case of deterministic service time itequals 0 For simplicity 1198882119904 = 05 Thus (17) can be rewrittenas

R asymp 34119879 (18)

Themean delay time for the NRT packet can be estimatedas

119882119899119903119905 = 1198821199031199051 minus 120601 (19)

The sum of gaps between the sequential transmissions isconsidered aswasted time and it is denoted by119882119905 119882119905 includesthe contention periods in which the channel is not chosen forsubmission Thus

119882119905 = 119875119894max119865

119899sum119895=1

10038161003816100381610038161003816119879119888119895 minus 11987911990111989510038161003816100381610038161003816 + (1 minus 119875119894) lfloor119879119900119891119891119879119888 rfloor119879119888 (20)

where 119875119894 is the probability that the BS will choose thischannel for the next submission The maximum number oftransmissions that can be achieved in the channelrsquos idle time119879119900119891119891 after deducting the time in which the channel was notchosen for the submission 119879119888119898 is given by

max119865119899 = 119879119900119891119891 minus 119879119888119898max (119879119888 119879119901) (21)

Accordingly (20) can be rewritten as

119882119905 asymp 119875119894max119865119899 10038161003816100381610038161003816119879119888119895 minus 11987911990111989510038161003816100381610038161003816 + (1 minus 119875119894) 119879119900119891119891 (22)

Each channel has the same probability to be chosen among119872 available channels since a uniform distribution is used toprovide the fairness property for the data channels thus theprobability of allocating a channel 119875119894 equals 1119872 Conse-quently the available time for data transmission equals (119879119900119891119891minus119882119905) and the achievable number of packet transmissions isgiven as

119873119879119901 = lfloor119879119900119891119891 minus 119882119905119879119901 rfloor (23)

The net transmission time is estimated as

119873119879119901 (119879119901 minus 119878119868119865119878 minus 119879119886119888119896) = 119873119879119901119871 (24)

Therefore the achieved throughput can be estimated as

Ψ119900119891119891 = 119873119879119901119871119863119889119903 (25)


Table 1 Evaluation parameters values

Parameter ValueContentionWindow min size 119862119882119898119894119899 32ContentionWindow max size 119862119882119898119886119909 1024CC Data Rate 119862119889119903(Mbps) 1Data Channel Data Rate119863119889119903(Mbps) 1Slot time in CC Channel (ms) 002119862119862119860 (ms) 002119863119868119865119878 (ms) 005119878119868119865119878 (ms) 001119879119903119905119904 (ms) 0352119879119888119905119904 (ms) 0304119879119886119888119896 (ms) 0304

Taking into account the probability that a node cantransmit in a specific idle time 120579119904 and the CC channelrsquosblocking probability 120579119887 that can be expressed as

120579119904 = expminus(119873119872)

and 120579119887 = 1 minus 119879119888119879(26)

(25) can be rewritten as

Ψ119900119891119891 = (1 minus 120579119904120579119887)119873119879119901119871119863119889119903 (27)

5 Packet Size Impact Evaluation

The evaluation of the impact based on different data packetsize was performed using MATLAB The main evaluationparameters values are listed in Table 1 and they are similarto that defined in IEEE 80211 MAC The duty cycle time isassumed to be 119900119899119890 second

Figure 3 shows the effects of changing the packet sizeon the relationship between the node population and theinduced delay for the two categories of the packets RT andNRT In the simulated scenario the channel idle time isfixed 119879119900119891119891 = 05 s number of the channels is 10 and thearrival rate at both of the queues is the same 120582119903119905 = 120582119899119903119905 =3 pkts The performance shows linear behavior of the RTpackets starting from the low to the high node densities whilethe performance of the NRT packets behaves exponentiallystarting from themediumnodes population119873 = 5 From thefigure the average delay increases as the packet size increaseseven though the delay values are similar in low nodes densitywith values less than 02 s The difference of the induceddelay values between RT and NRT packets increases as nodesdensity increases and that can be explained given that the RTpackets have the advantage in the CC contention

In Figure 4 the trend towards the exponential behaviorfor both RT and NRT packet delay curves is so clear Inthis simulation the node population is fixed at 119873 = 5 3pkts is the arrival rate for each node and 10 channels arebeing utilised In short busy time periods the values of theaverage delay are low below 01 s As the channelrsquos busy timeincreases the delay increases for all packet sizes although

1 2 3 4 5 6 7 8 9 10Nodes density

0

02

04

06

08

1

12

Aver

age d

elay

tim

e (se

cond

)

RT L=1kbNRT L=1kbRT L=3kb

NRT L=3kbRT L=10kbNRT L=10kb

Figure 3 Average delay vs nodes density

01 02 03 04 05 06 07 08 09Channel busy time length

0

01

02

03

04

05

06

07

08

09

1

Aver

age d

elay

tim

e (se

cond

)



046 048 05 052

01

02

03

Figure 4 Average delay vs channelrsquos busy time

the large-size packets are exposed to longer delay When thepacket size is 10 kb and 119879119900119899 = 09 s the induced delay recordsits highest values 043 s and 094 s for RT and NRT packetsrespectively

Figure 5 presents the effects of the packet size on therelationship between the number of the available channelsand the average delay Nodes density is fixed at 119873 = 5and channel idle time is 119879119900119899 = 05 s The induced delay forNRT packets is very long when the number of the availablechannels is less than 5 to the extent that the system losses its


4 5 6 7 8 9 10Number of channels

0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)



Figure 5 Average delay vs number of channels

stability with large-size packets ie 119871 = 10 kb The curves ofthe induced delay fall steeply for NRT packets as the numberof channels increases to 5 channels and gradually decline as119872 exceeds 5 channels For large-size packets of 10 kb thedeclining trend is slower compared to other packet sizesFrom the figure it is obvious thatmedium-size packets of 3 kbcan cope with the miserly licensed radio environment whilelarge-size packets cannot Resulting interference contributesmore to increasing the induced delay for large-size packetssince the collided packet has to be retransmitted again

Figure 6 shows the performances of the PCMACprotocoland DOSS MAC protocol with various sizes of the datapacket The latter protocol was improved and modelled in[17] Nodes density equals 5 and 7 data channels are exploitedAlthough DOSS MAC offers a shorter delay time in shortlengths of channels busy times it behaves exponentially inlonger busy times starting from 04 s PCMAC presents amore stable performance in both short and long lengths ofchannel idle times Generally the average delay time increasesas the packet size becomes larger in both of PCMAC andDOSS protocols

Figure 7 illustrates that the throughput can be signifi-cantly affected by packet size In the simulation the idle timeand the number of channels are fixed at 119879119900119891119891 = 05 and119872 = 5 respectively The throughput curves growths havean exponential trend for all packet sizes however the curveof small-size packet 119871 = 1 kb shows the slowest trend Theimpact of the time gap between the contention time 119879119888 andthe packet size 119871 is clear in case of 119871 = 1 kb As the number ofnodes increases the contention time increases ie the timegap between119879119901 and119879119888 decreases until it approaches the valueof the packet size at which the throughput reaches its peak inthe range of 119873 = 7 and 119873 = 8 before declining gradually


0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)

PCMAC L=1 kbDOSS L=1 kb


Figure 6 Impact of the packet size on the induced delay in PCMAC[16] and DOSS [17] protocols

1 2 3 4 5 6 7 8 9Number of nodes

35

4

45

5

55

6

65

7

75

8

85

Thro

ughp

ut (b

itss

econ

d)

L= 1 kbL= 3 kbL= 10 kb

times104

Figure 7Throughput vs nodes density

In this simulation medium-size packets show a much closerperformance to that of large-size packets

In Figure 8 the significant impact of the packet sizeon the relationship between the channel idle time and theachieved throughput is illustrated Generally the throughputincreases dramatically as channel idle time increases butwith packets of medium size 119871 = 3 kb the throughputperformance is better compared to other sizes For packetsof 1 kb size the time gap between the contention timeand the packet time plays a main role since the packet


01 02 03 04 05 06 07 08 09Channel idle time length

0

2

4

6

8

10

12

14

Thro

ughp

ut (b

itss

econ

d)

L=1 kbL=3 kbL=10 kb

times104

Figure 8 Throughput vs channelrsquos idle time

size determines throughput performance for larger packetsThe higher the achieved throughput the more efficient theutilization of the radio channels The fluctuation of thethroughput curve for packets of 10 kb size is because ofthe submission nature in which the receiver deals with thedata packet as a block As a matter of fact the latter featureis also responsible for the high probability of interferenceleading to longer delay for large-size packets Because ofthe burst nature of the traffic in PCMAC protocol it showsbetter performance than DOSSMACprotocol in terms of theachievable throughput especially with small-size packets asillustrated in Figure 9 Using medium-size packets 119871 = 3 kbimproves DOSS MAC protocolrsquos performance significantlyalthough the nodes density increasing will result in slowingdown the throughput increasing due to the contention inthe CC channel The simultaneous transmission natureie concurrent submission through many channels for thepackets inDOSS causes the high sensitivity against increasingnodes density

In general at the same time the medium-size packetsoutperform small-size packets in terms of the achievedthroughput and channelrsquos utilization they show closer perfor-mances to them in terms of the latency Moreover employingpackets of size 375 byte can be the optimal choice whenusing an adaptive architecture for the network such as thatproposed in [26] Packets of medium size that are closer tothe common sizes used in WSN standards up to 128 bytesensure a robust performance in the alternative workflow thatis CR-based

6 Conclusion

The size of the data packet in CRSN networks plays a keyrole not only to enhance the throughput and increase the


0

05

1

15

2

25

3

35

4

Agg

rega

ted

thro

ughp

ut (b

itss

econ

d)

PCMAC L=1kbDOSS L=1kb


times105

Figure 9 Achieved throughput comparison between PCMAC [16]and DOSS [17]

efficiency of channelsrsquo utilization but also to mitigate theharmful interference This study has examined the impact ofpacket size on the performance of CRSN in terms of twomainperformance metrics namely the delay and the throughputSimulation results show that exploiting the appropriate sizewithin an efficient MAC protocol such as PCMAC sig-nificantly enhances the performance Medium-size packetsoutperformance is proven by the short induced latency whichis suitable for critical data and the increased throughputFurthermore the results show that large-size packets notonly fail to cope with a poor radio environment but alsodo not enhance the throughput significantly in contrast tothe packets of medium size that perform better in severalconditions Through this study it is evident that medium-sizepackets are the optimal choice for adaptive CRSN networksdespite the current trends of using small-size packets

Data Availability

No data were used to support this study

Conflicts of Interest

The authors declare that they have no conflicts of interest

References

[1] S Farahani ldquoZigbee basicsrdquo in ZigBee Wireless Networks andTransceivers Elsevier 1st edition 2008

[2] M Becker Services in Wireless Sensor Networks Modelling andOptimisation for the Efficient Discovery of Services SpringerScience amp Business Media 2014

[3] D Yang Y Xu andM Gidlund ldquoWireless coexistence betweenIEEE 80211- and IEEE 802154-based networks a surveyrdquo


International Journal of Distributed Sensor Networks vol 2011Article ID 912152 17 pages 2011

[4] Federal Communications Commission (FCC) Docket no 03-222 notice of proposed rule making and order 2003

[5] A M Wyglinski M Nekovee and Y T Hou Cognitive RadioCommunications and Networks Principles and Practice Aca-demic Press 2009

[6] D Cabric S MMishra and RW Brodersen ldquoImplementationissues in spectrum sensing for cognitive radiosrdquo in Proceedingsof the Conference Record of the 38th Asilomar Conference onSignals Systems and Computers pp 772ndash776 Pacific GroveCalif USA November 2004

[7] W-Y Lee and I F Akyildiz ldquoOptimal spectrum sensingframework for cognitive radio networksrdquo IEEE Transactions onWireless Communications vol 7 no 10 pp 3845ndash3857 2008

[8] G P Joshi S Y Nam and S W Kim ldquoCognitive radio wirelesssensor networks applications challenges and research trendsrdquoSensors vol 13 no 9 pp 11196ndash11228 2013

[9] J Ren Y Zhang N Zhang D Zhang et al ldquoDynamic channelaccess for energy efficient data gathering in cognitive radiosensor networksrdquo httpsarxivorgabs150706188

[10] Zilong Jin Yu Qiao Alex Liu and Lejun Zhang ldquoEESS AnEnergy-Efficient Spectrum Sensing Method by OptimizingSpectrum Sensing Node in Cognitive Radio Sensor NetworksrdquoWireless Communications and Mobile Computing vol 2018Article ID 9469106 11 pages 2018

[11] A Sawabe K Tsukamoto and Y Oie ldquoQoS-aware packetchunking schemes for M2M cloud servicesrdquo in Proceedingsof the 28th IEEE International Conference on Advanced Infor-mation Networking and Applications Workshops IEEE WAINA2014 pp 166ndash173 Canada May 2014

[12] L Song D Niyato Z Han and E Hossain Wireless Device-to-Device Communications and Networks Cambridge UniversityPress Cambridge UK 2015

[13] N U Hasan W Ejaz and H S Kim ldquoFrame size selectionin CSMA-based cognitive radio wireless local area networksrdquoTransactions on Emerging Telecommunications Technologies vol28 no 1 2017

[14] G Durisi T Koch and P Popovski ldquoTowards massive ultra-reliable and low-latency wireless communication with shortpacketsrdquo httpsarxivorgabs150406526

[15] S Atapattu C Tellambura and H Jiang ldquoConventional EnergyDetectorrdquo in Energy Detection for Spectrum Sensing in CognitiveRadio pp 11ndash26 Springer New York NY USA 2014

[16] M Al-Medhwahi F Hashim B M Ali and A Sali ldquoPliablecognitive MAC for heterogeneous adaptive cognitive radiosensor networksrdquo PLoS ONE vol 11 no 6 p e0156880 2016

[17] G A Shah and O B Akan ldquoPerformance analysis of CSMA-based opportunistic medium access protocol in cognitive radiosensor networksrdquo Ad Hoc Networks vol 15 pp 4ndash13 2014

[18] O Ergul andO B Akan ldquoCooperative coarse spectrum sensingfor cognitive radio sensor networksrdquo in Proceedings of the 2014IEEE Wireless Communications and Networking ConferenceWCNC 2014 pp 2055ndash2060 Turkey April 2014

[19] G A Shah and O B Akan ldquoCognitive adaptive medium accesscontrol in cognitive radio sensor networksrdquo IEEE Transactionson Vehicular Technology vol 64 no 2 pp 757ndash767 2015

[20] L Ma X Han and C-C Shen ldquoDynamic open spectrum shar-ing MAC protocol for wireless ad hoc networksrdquo in Proceedingsof the 2005 1st IEEE International Symposium on New Frontiersin Dynamic Spectrum Access Networks DySPAN 2005 pp 203ndash213 USA November 2005

[21] W Tang M Z Shakir M A Imran R Tafazolli and M-S Alouini ldquoThroughput analysis for cognitive radio networkswith multiple primary users and imperfect spectrum sensingrdquoIET Communications vol 6 no 17 pp 2787ndash2795 2012

[22] N U Hasan W Ejaz and H S Kim ldquoFrame size selectionin CSMA-based cognitive radio wireless local area networksrdquoTransactions on Emerging Telecommunications Technologies2014

[23] T O Kim A S Alfa and B D Choi ldquoPerformance Analysisof a CSMACA Based MAC Protocol for Cognitive RadioNetworksrdquo in Proceedings of the 2010 IEEE Vehicular TechnologyConference (VTC 2010-Fall) pp 1ndash5 Ottawa ON CanadaSeptember 2010

[24] K J Kim K S Kwak and B D Choi ldquoPerformance analysisof opportunistic spectrum access protocol for multi-channelcognitive radio networksrdquo Journal of Communications andNetworks vol 15 no 1 pp 77ndash86 2013

[25] BGulbahar andO B Akan ldquoInformation theoretical optimiza-tion gains in energy adaptive data gathering and relaying incognitive radio sensor networksrdquo IEEE Transactions onWirelessCommunications vol 11 no 5 pp 1788ndash1796 2012

[26] M Al-Medhwahi and F Hashim ldquoThe adaptive CognitiveRadio Sensor Network A perspective towards the feasibilityrdquo inProceedings of the 1st International Conference on Telematics andFuture Generation Networks TAFGEN 2015 pp 6ndash11 MalaysiaMay 2015

[27] IEEE 802 Working Group ldquoIeee standard for local andmetropolitan area networksrdquo part 154 Low-rate wireless per-sonal area networks (lr-wpans)rdquo IEEE Std vol 802 pp 4ndash20112011

[28] P Huang L Xiao S Soltani M W Mutka and N Xi ldquoTheevolution of MAC protocols in wireless sensor networks asurveyrdquo IEEE Communications Surveys amp Tutorials vol 15 no1 pp 101ndash120 2013

[29] IEEE Standards Association ldquoPart 22 Cognitive wireless ranmedium access control (mac) and physical layer (phy) specifi-cations Policies and procedures for operation in the tv bandsrdquoIEEE Standard vol 802 2011

[30] H Zhao E Garcia-Palacios J Wei and Y Xi ldquoAccurateavailable bandwidth estimation in IEEE 80211-based ad hocnetworksrdquo Computer Communications vol 32 no 6 pp 1050ndash1057 2009

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018


Active and Passive Electronic Components

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of



Journal ofEngineeringVolume 2018

SensorsJournal of



RotatingMachinery


Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation


Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom


PU Base Station

CRSN Base Station

PU User

Sensor

SU User

Figure 1 Typical cognitive radio sensor network

where 119882(119899) represents the additive white Gaussian noise(AWGN) and 119878(119899) represents the transmitted signal multi-plied by the channel gain

The probability that the PU channel is busy is given by

119875119900119899 = 119879119900119899119879119900119899 + 119879119900119891119891 (2)

with 119879119900119899 as the channel busy time and 119879119900119891119891 as channel idletime The probability that the channel is free 119875119900119891119891 is given as

119875119900119891119891 = 119879119900119891119891119879119900119899 + 119879119900119891119891 (3)

Probability that the channel is detected as busy when itis actually busy is called the detection probability 119875119889119890 andprobability that the channel is detected as busy when it is idleis called the false alarm probability 119875119891119886 Optimal detectioncan be achieved only if the noise power is known to the SUsThe channel status is determined based on the value of thereceived signalrsquos energy119864 and a predefined threshold value119879119903 The relationship between 119875119889119890 and 119875119891119886 with 119879119903 and 119864 is givenby (4) and (5) respectively

119875119889119890 = 119875119903 119863 = 1 | 1198671 = 119875119903 119864 gt 119879119903 | 1198671 (4)

and

119875119891119886 = 119875119903 119863 = 1 | 1198670 = 119875119903 119864 gt 119879119903 | 1198670 (5)

Emerging technology that enables the opportunisticaccess for WSNrsquos units is called the cognitive radio sensornetwork (CRSN) whose typical layout is shown in Figure 1Appropriate data packet size helps limit the interferencebetween the signals of the WSN units equipped with CRcapabilities and PU signals [8ndash10] Thus size of data packetin CRSN has a twofold importance to achieve a satisfactoryperformance and to mitigate harmful interference with PUsignal Small data packets offer maximum reliability andminimum latency for most types of data traffic especially incritical traffic Together with the growth of sensor network ininternet of things (IoT) the result is that small data packetsof sizes up to 128 bytes (1024 bits) have become the populartrend in data transmission [11 12] Subsequently overheaddata and the inefficient utilisation of resources induced byusing small-size packets in emerging heterogeneous networksare often overlooked Other substantial factors such as small-size packets have low signal noise ratios (SNRs) compared tolarge size packets in effect of noise such as thermal noise [13]and the fact that large-size packets are able to achieve higherefficiency of bandwidth utilisation [14 15] is neglected

The contributions of this study can be summarised asfollows

(i) It investigates the impact of packet size on the per-formance of CRSNs in terms of two main metricsnamely the average delay and the throughput





2 Related Works



3 System Model





Channel

SU 1

SU 2


Submitted Packet

Failed Packet

Spectrum Sensing

PU (On)PU (Off)








119879119905119903 = 119879119888 + 119879119904119904 + 119879119901 (6)





119879119904119904 = 1198791199001198992 119875119900119899 (8)



119879119901 = 119862119862119860 + 119871119863119889119903 + 119878119868119865119878 + 119879119886119888119896 (9)


where




120601119903119905119888 = 120582119903119905119879119888and 120601119899119903119905119888 = 120582119899119903119905119888119879119888 (10)


120601119903119905119889 = 119873120582119903119905119879119904119904 + 119879119901119862and 120601119899119903119905119889 = 119873120582119899119903119905119879119904119904 + 119879119901119862

(11)


119882119903119905 = 119871119903119905119879 + ΦR (12)


R = 11987922119879 = 1205902119879 + 1198792

2119879 (13)


119882119903119905 (1 minus 120601119903119905) = 120601119903119905R (14)


119882119903119905 = ΦR1 minus 120601119903119905 (15)


119879119903119905 = 119882119903119905 + 119879 (16)


R = 12 (1198882119904 + 1) 119879 (17)


R asymp 34119879 (18)


119882119899119903119905 = 1198821199031199051 minus 120601 (19)


119882119905 = 119875119894max119865

119899sum119895=1



max119865119899 = 119879119900119891119891 minus 119879119888119898max (119879119888 119879119901) (21)






119873119879119901 (119879119901 minus 119878119868119865119878 minus 119879119886119888119896) = 119873119879119901119871 (24)


Ψ119900119891119891 = 119873119879119901119871119863119889119903 (25)





120579119904 = expminus(119873119872)

and 120579119887 = 1 minus 119879119888119879(26)


Ψ119900119891119891 = (1 minus 120579119904120579119887)119873119879119901119871119863119889119903 (27)





1 2 3 4 5 6 7 8 9 10Nodes density

0

02

04

06

08

1

12

Aver

age d

elay

tim

e (se

cond

)





0

01

02

03

04

05

06

07

08

09

1

Aver

age d

elay

tim

e (se

cond

)



046 048 05 052

01

02

03






0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)








0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)





35

4

45

5

55

6

65

7

75

8

85

Thro

ughp

ut (b

itss

econ

d)


times104






0

2

4

6

8

10

12

14

Thro

ughp

ut (b

itss

econ

d)

L=1 kbL=3 kbL=10 kb

times104




6 Conclusion



0

05

1

15

2

25

3

35

4

Agg

rega

ted

thro

ughp

ut (b

itss

econ

d)



times105



Data Availability




References



































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia






2 Related Works



3 System Model





Channel

SU 1

SU 2


Submitted Packet

Failed Packet

Spectrum Sensing

PU (On)PU (Off)








119879119905119903 = 119879119888 + 119879119904119904 + 119879119901 (6)





119879119904119904 = 1198791199001198992 119875119900119899 (8)



119879119901 = 119862119862119860 + 119871119863119889119903 + 119878119868119865119878 + 119879119886119888119896 (9)


where




120601119903119905119888 = 120582119903119905119879119888and 120601119899119903119905119888 = 120582119899119903119905119888119879119888 (10)


120601119903119905119889 = 119873120582119903119905119879119904119904 + 119879119901119862and 120601119899119903119905119889 = 119873120582119899119903119905119879119904119904 + 119879119901119862

(11)


119882119903119905 = 119871119903119905119879 + ΦR (12)


R = 11987922119879 = 1205902119879 + 1198792

2119879 (13)


119882119903119905 (1 minus 120601119903119905) = 120601119903119905R (14)


119882119903119905 = ΦR1 minus 120601119903119905 (15)


119879119903119905 = 119882119903119905 + 119879 (16)


R = 12 (1198882119904 + 1) 119879 (17)


R asymp 34119879 (18)


119882119899119903119905 = 1198821199031199051 minus 120601 (19)


119882119905 = 119875119894max119865

119899sum119895=1



max119865119899 = 119879119900119891119891 minus 119879119888119898max (119879119888 119879119901) (21)






119873119879119901 (119879119901 minus 119878119868119865119878 minus 119879119886119888119896) = 119873119879119901119871 (24)


Ψ119900119891119891 = 119873119879119901119871119863119889119903 (25)





120579119904 = expminus(119873119872)

and 120579119887 = 1 minus 119879119888119879(26)


Ψ119900119891119891 = (1 minus 120579119904120579119887)119873119879119901119871119863119889119903 (27)





1 2 3 4 5 6 7 8 9 10Nodes density

0

02

04

06

08

1

12

Aver

age d

elay

tim

e (se

cond

)





0

01

02

03

04

05

06

07

08

09

1

Aver

age d

elay

tim

e (se

cond

)



046 048 05 052

01

02

03






0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)








0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)





35

4

45

5

55

6

65

7

75

8

85

Thro

ughp

ut (b

itss

econ

d)


times104






0

2

4

6

8

10

12

14

Thro

ughp

ut (b

itss

econ

d)

L=1 kbL=3 kbL=10 kb

times104




6 Conclusion



0

05

1

15

2

25

3

35

4

Agg

rega

ted

thro

ughp

ut (b

itss

econ

d)



times105



Data Availability




References



































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



Channel

SU 1

SU 2


Submitted Packet

Failed Packet

Spectrum Sensing

PU (On)PU (Off)








119879119905119903 = 119879119888 + 119879119904119904 + 119879119901 (6)





119879119904119904 = 1198791199001198992 119875119900119899 (8)



119879119901 = 119862119862119860 + 119871119863119889119903 + 119878119868119865119878 + 119879119886119888119896 (9)


where




120601119903119905119888 = 120582119903119905119879119888and 120601119899119903119905119888 = 120582119899119903119905119888119879119888 (10)


120601119903119905119889 = 119873120582119903119905119879119904119904 + 119879119901119862and 120601119899119903119905119889 = 119873120582119899119903119905119879119904119904 + 119879119901119862

(11)


119882119903119905 = 119871119903119905119879 + ΦR (12)


R = 11987922119879 = 1205902119879 + 1198792

2119879 (13)


119882119903119905 (1 minus 120601119903119905) = 120601119903119905R (14)


119882119903119905 = ΦR1 minus 120601119903119905 (15)


119879119903119905 = 119882119903119905 + 119879 (16)


R = 12 (1198882119904 + 1) 119879 (17)


R asymp 34119879 (18)


119882119899119903119905 = 1198821199031199051 minus 120601 (19)


119882119905 = 119875119894max119865

119899sum119895=1



max119865119899 = 119879119900119891119891 minus 119879119888119898max (119879119888 119879119901) (21)






119873119879119901 (119879119901 minus 119878119868119865119878 minus 119879119886119888119896) = 119873119879119901119871 (24)


Ψ119900119891119891 = 119873119879119901119871119863119889119903 (25)





120579119904 = expminus(119873119872)

and 120579119887 = 1 minus 119879119888119879(26)


Ψ119900119891119891 = (1 minus 120579119904120579119887)119873119879119901119871119863119889119903 (27)





1 2 3 4 5 6 7 8 9 10Nodes density

0

02

04

06

08

1

12

Aver

age d

elay

tim

e (se

cond

)





0

01

02

03

04

05

06

07

08

09

1

Aver

age d

elay

tim

e (se

cond

)



046 048 05 052

01

02

03






0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)








0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)





35

4

45

5

55

6

65

7

75

8

85

Thro

ughp

ut (b

itss

econ

d)


times104






0

2

4

6

8

10

12

14

Thro

ughp

ut (b

itss

econ

d)

L=1 kbL=3 kbL=10 kb

times104




6 Conclusion



0

05

1

15

2

25

3

35

4

Agg

rega

ted

thro

ughp

ut (b

itss

econ

d)



times105



Data Availability




References



































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



where




120601119903119905119888 = 120582119903119905119879119888and 120601119899119903119905119888 = 120582119899119903119905119888119879119888 (10)


120601119903119905119889 = 119873120582119903119905119879119904119904 + 119879119901119862and 120601119899119903119905119889 = 119873120582119899119903119905119879119904119904 + 119879119901119862

(11)


119882119903119905 = 119871119903119905119879 + ΦR (12)


R = 11987922119879 = 1205902119879 + 1198792

2119879 (13)


119882119903119905 (1 minus 120601119903119905) = 120601119903119905R (14)


119882119903119905 = ΦR1 minus 120601119903119905 (15)


119879119903119905 = 119882119903119905 + 119879 (16)


R = 12 (1198882119904 + 1) 119879 (17)


R asymp 34119879 (18)


119882119899119903119905 = 1198821199031199051 minus 120601 (19)


119882119905 = 119875119894max119865

119899sum119895=1



max119865119899 = 119879119900119891119891 minus 119879119888119898max (119879119888 119879119901) (21)






119873119879119901 (119879119901 minus 119878119868119865119878 minus 119879119886119888119896) = 119873119879119901119871 (24)


Ψ119900119891119891 = 119873119879119901119871119863119889119903 (25)





120579119904 = expminus(119873119872)

and 120579119887 = 1 minus 119879119888119879(26)


Ψ119900119891119891 = (1 minus 120579119904120579119887)119873119879119901119871119863119889119903 (27)





1 2 3 4 5 6 7 8 9 10Nodes density

0

02

04

06

08

1

12

Aver

age d

elay

tim

e (se

cond

)





0

01

02

03

04

05

06

07

08

09

1

Aver

age d

elay

tim

e (se

cond

)



046 048 05 052

01

02

03






0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)








0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)





35

4

45

5

55

6

65

7

75

8

85

Thro

ughp

ut (b

itss

econ

d)


times104






0

2

4

6

8

10

12

14

Thro

ughp

ut (b

itss

econ

d)

L=1 kbL=3 kbL=10 kb

times104




6 Conclusion



0

05

1

15

2

25

3

35

4

Agg

rega

ted

thro

ughp

ut (b

itss

econ

d)



times105



Data Availability




References



































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia






120579119904 = expminus(119873119872)

and 120579119887 = 1 minus 119879119888119879(26)


Ψ119900119891119891 = (1 minus 120579119904120579119887)119873119879119901119871119863119889119903 (27)





1 2 3 4 5 6 7 8 9 10Nodes density

0

02

04

06

08

1

12

Aver

age d

elay

tim

e (se

cond

)





0

01

02

03

04

05

06

07

08

09

1

Aver

age d

elay

tim

e (se

cond

)



046 048 05 052

01

02

03






0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)








0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)





35

4

45

5

55

6

65

7

75

8

85

Thro

ughp

ut (b

itss

econ

d)


times104






0

2

4

6

8

10

12

14

Thro

ughp

ut (b

itss

econ

d)

L=1 kbL=3 kbL=10 kb

times104




6 Conclusion



0

05

1

15

2

25

3

35

4

Agg

rega

ted

thro

ughp

ut (b

itss

econ

d)



times105



Data Availability




References



































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)








0

02

04

06

08

1

12

14

16

18

2

Aver

age d

elay

tim

e (se

cond

)





35

4

45

5

55

6

65

7

75

8

85

Thro

ughp

ut (b

itss

econ

d)


times104






0

2

4

6

8

10

12

14

Thro

ughp

ut (b

itss

econ

d)

L=1 kbL=3 kbL=10 kb

times104




6 Conclusion



0

05

1

15

2

25

3

35

4

Agg

rega

ted

thro

ughp

ut (b

itss

econ

d)



times105



Data Availability




References



































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




0

2

4

6

8

10

12

14

Thro

ughp

ut (b

itss

econ

d)

L=1 kbL=3 kbL=10 kb

times104




6 Conclusion



0

05

1

15

2

25

3

35

4

Agg

rega

ted

thro

ughp

ut (b

itss

econ

d)



times105



Data Availability




References



































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia

































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


Impact of Packet Size in Adaptive Cognitive Radio Sensor ...

Documents

Transcript of Impact of Packet Size in Adaptive Cognitive Radio Sensor ...