ASMS2004 Tutorials DVB-S2 - ESA ARTES · PDF fileASMS2004 Tutorials DVB-S2 Giovanni E. Corazza...

1

ASMS2004 TutorialsASMS2004 Tutorials

DVBDVB--S2S2Giovanni E. Giovanni E. CorazzaCorazza

NoordwijkNoordwijk, Sep 20, 2004, Sep 20, 2004

2EMPS/ASMS 2004 Conference – ESTEC Noordwijk, September 20, 2004

Motivation for DVBMotivation for DVB--S2S2

DVB-S2 is the second generation DVB system for broadband satellite servicesDRAFT ETSI EN 302 307 (2004-01)The DVB-S legacy:

DVB-S is an extremely successful standardDVB-S was devised for broadcasting applications where

the physical layer is fixed and optimized for the worst-case

For unicast applications this worst-case design approach causes a large waste of satellite resources

DVB-S2 is designed to increase spectral efficiency and qualityDVB-S2 does so by combining:

Adaptive Coding and Modulation (ACM)Near Shannon-limit Forward Error Correction (FEC)

Multi-spot Ka-Band satellite antennaNew video and audio coding schemes can be used


DVBDVB--S2 ApplicationsS2 Applications

Broadcast Services (BS) BS as for DVB-S, but with the added flexibility of VCM (Variable Coding and Modulation) enabling different levels of protection for each service (e.g. robust SDTV, with less-robust HDTV)BC-BS (backwards compatible broadcast services) for interoperability with DVB-S decoders, and a more optimised NBC-BS (non-backwards compatible)

Interactive Services (IS) IS are provided with existing DVB return channel standards (e.g. RC-PSTN, RCS, etc.)

DVB-S2 can operate in CCM (constant coding & modulation) and ACM (Adaptive Coding and Modulation) modesDigital TV Contribution and Satellite News Gathering (DTVC/DSNG)DTVC/DSNG builds on the DVB-DSNG standard, facilitating point-to-point, or point-to-multipoint communications of single or multiple MPEG transport streams using either CCM, or ACM modes.

Professional Applications (PS) E.g. data content distribution/trunking, using CCM, VCM or ACM

2


ACM BackgroundACM Background

INTERFERENCE DISTRIBUTION:ACM exploits the entire C/I range ATMOSPHERIC CONDITIONS:

ACM maximizes instantaneous data-rateas a function of time/location


How does ACM operate?How does ACM operate?

According to the user SNR, the system selects the optimum couple [code-rate / modulation]

16APSK

QPSK8PSK


Shannon bound for MShannon bound for M--ary modulationsary modulations

Modulation

Scheme Losses from

Shannon Bound QPSK 1.75 dB 8PSK 2.2 dB

16QAM 4.1 dB 32QAM 6.1 dB 64QAM 7.5 dB

3


Discretization for coded MDiscretization for coded M--ary modulationsary modulations


DVBDVB--S2: Tx Block DiagramS2: Tx Block Diagram


Mode AdaptationMode Adaptation

MA (Mode Adaptation) = describe the input stream

UPL (User Packet Length) = constant value, 188bytes (MPEG)

UPL = Number of UP bits, [0, 64k]

DFL = Data field length in bits

SYNC = Synchronization byte

SYNCD = Distance in bits from the first UP

CRC applied to BBHEADER

4


FEC FEC encodingencoding

Encoding is performed in three stages:outer coding �� BCH

Parity check bits (BCHFEC) of BCH outer code are appended to BBFRAME

Inner Coding �� LDPC

Parity check bits (LDPCFEC) of LDPC encoder are appended to BCHFEC field

Bit interleaving

Each BBFRAME (KBCH bits) are processed by the FEC coding subsystem to generate a FECFRAME �� NLDPC bits


Coding parametersCoding parameters

Two different FECFRAME formats are available:The “normal” FECFRAME �� NLDPC = 64800The “short” FECFRAME �� NLDPC = 16200

Available code-rates: 1/4, 1/3, 2/5, 1/2, 3/5, 2/3, 3/4, 4/5, 5/6, 8/9, 9/10


LowLow--density paritydensity parity--check codescheck codes

1 1 1 11 1 1 1

1 1 1 11 1 1 1

1 1 1 11 1 1 1

1 1 1 11 1 1 1

1 1 1 11 1 1 1

1 1 1 11 1 1 1

Defined also in terms of sparse random graphsTheir property is that each constraint involves a small number of variables in the graphThe number of edges in the graph scales roughly linearly with N, rather than quadraticallyRandomness ensures a good code while sparseness enable efficient decoding

Decoded by a simple probability-based message-passing algorithmMost used is the sum-product algorithm

Is not the optimal decoderResults are record-breaking

Discovered by Gallager (1961)Rediscovered during 1990’s by MackayBlock codes that have parity-check matrix H every row and column of which is “sparse”

Regular LDPCEvery column of H has the same weight jand every row as the same weight k

Irregular LDPCEvery column and row of H has a particular bit-node and check-node degree

5


Message PassingMessage Passing

Practical codes should have simple coding and decoding, but generally the decoding requires a lot time and resourcesThis problem can be solved by message-passing algorithm

Complicated calculations are simple distributed among simple processorsAfter a few steps the solution of the global problem is available

Commander

1234

4321


Message PassingMessage Passing

This algorithm can be used if the “soldiers” are arranged in a graph that contains no cycles!If there are cycles, a modified message-passing algorithm must be used

Cycles are opened and a cycle-free graph (tree) is obtained

Commander


SumSum--product algorithm: the most likely pathproduct algorithm: the most likely path

B

1 1A 1 1

1 2 3 4

2 5 9

5 14

14

1

1

1

1

1

3

4

5

9

5

14

1

14

2 2

1

B

1 1 1 1

1 2 3 4

2 5 9

5 14

14

1A

The sum takes place at each node by adding messages from predecessors

The product comes in by weighting terms in the sum

The sum-product algorithm is a form of belief propagation

Forwardrecursion

Backwardrecursion

6


Code Construction and DecodingCode Construction and Decoding

There is a relationship between the locations of 1’s in the matrix and

the cycles in the code graph the number of decoding iterations

Good codes should have graphs without short cycles (closed loops)

A closed loop in parity-check matrix is a sequence of connected alternating horizontal and vertical lines The last lines in the sequence terminates at the beginning of the first lineEvery line stats in a vertex, which is a point where the parity check matrix contains a 1

The sum-product decoding algorithm is optimal if there are no cycles ��sub-optimal with cyclesIncreasing the matrix size, it becomes easy to produce matrices without cycles of any given length �� it is possible to ignore cycles

1 1 1 11 1 1 1

1 1 1 11 1 1 1

1 1 1 11 1 1 1

1 1 1 11 1 1 1

1 1 1 11 1 1 1

1 1 1 11 1 1 1


DVBDVB--S2: BitS2: Bit--InterleaverInterleaver

Performed only for 8-PSK, 16-APSK, and 32-APSK modulation schemesLDPC encoder output is interleaved using a simple block interleaver

Interleaving depth is a function of the adopted modulation format

5 * 1296032-PSK

4 * 1620016-PSK

3 * 216008-PSK

SizeModulation

Bit-Interleaver Structure


Modulation SchemesModulation Schemes

LDPC codewords are mapped into the following constellations:

QPSK 8-PSK

16-APSK 32-APSK

7


ACMACM Performance: AWGNPerformance: AWGN

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10

Eb/N0 [dB]

PE

R

1/2 - QPSK

3/5 - QPSK

2/3 - QPSK

3/4 - QPSK

2/3 - 8PSK

3/4 - 8PSK

5/6 - 8PSK

3/4 - 16APSK

5/6 - 16APSK

3/4 - 32APSK


ACMACM Performance: AWGN + NonPerformance: AWGN + Non--linearitieslinearities

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10 10.5 11 11.5 12

Eb/N0 [dB]

PE

R

1/2 - QPSK (IC)

1/2 - QPSK (IBO=2dB)

3/5 - QPSK (IC)


2/3 - QPSK (IC)


3/4 - QPSK (IC)


2/3 - 8PSK (IC)

2/3 - 8PSK (IBO=2dB)

3/4 - 8PSK (IC)


5/6 - 8PSK (IC)


3/4 - 16APSK (IC)

3/4 - 16APSK (IBO=2dB)

5/6 - 16APSK (IC)


3/4 - 32APSK (IC)


3/4 - 32APSK (IBO=3.5dB)


ACM Performance: AWGN + Residual Error Phase (RPE)

1.00E-06

1.00E-05

1.00E-04

1.00E-03

1.00E-02

1.00E-01

1.00E+00

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6

Eb/No

PE

R

Rc=1/2, QPSK, REP=0 Rc=2/3, 8PSK, REP=0 Rc=5/6, 8PSK, REP=0

Rc=1/2, QPSK, REP=7.0 Rc=2/3, 8PSK, REP=3.0 Rc=5/6, 8PSK, REP=5.0

8


Mode AdaptationMode Adaptation

The receiver estimates the instantaneous SNR

This is fed-back to the gateway station by the return channel (DVB-RCS)

The gateway adapts the transmission mode accordingly

SNR estimation is not a trivial task


SNORE: Analytical CharacterizationSNORE: Analytical Characterization

)(ˆ kγ

� ⋅ ) Re(1

pN ( )2⋅

2 ⋅ � ⋅ ) (1

pN

nR ÷)(ˆ kPR

)(ˆ kPS

)(ˆ kPN

*n

d )(ˆ kγ

� ⋅ ) Re(1

pN ( )2⋅

2 ⋅ � ⋅ ) (1

pN

nR ÷)(ˆ kPR

)(ˆ kPS

)(ˆ kPN

*n

d


SNORE: CramerSNORE: Cramer--Rao BoundRao Bound

CRB

Mean Estimation Error

NOISE ACTS BOTH AS NUISANCE AND TARGET!

9


DVBDVB--S2: more infoS2: more info

www.dvb.org

http://www.dvb.org/documents//en302307.v1.1.1.draft.pdf

http://www.dvb.org/documents/white-papers/wp06.DVB-S2.final.pdf

ASMS2004 TutorialsASMS2004 Tutorials

DVBDVB--RCSRCSGiovanni E. Giovanni E. CorazzaCorazza

NoordwijkNoordwijk, Sep 20, 2004, Sep 20, 2004


Interactivity in previously under-served locations Extend services in areas lacking conventional terrestrial infrastructure

Internet Service Providers (ISPs)Application Service Providers (ASPs)UMTS base-stationsDigital Subscriber Line (xDSL) nodesPoint-to-Multipoint (PMP) radio base-stations

DVBDVB--RCS RCS ServicesServices and and ApplicationsApplications (1/2)(1/2)

10


DVBDVB--RCS RCS ServicesServices and and ApplicationsApplications (2/2)(2/2)

For government, education and business:Real-Time services and applications

VoIP and VideoconferencingCoLo: Colocation for web servers and web hosting Finance and stock market services Banking and financial servicesLAN Interconnection: VPNsDistance learning

Video, text, voiceTele-medicineInteractive TV broadcastingDistributed TV Broadcast IP Multicast and IP StreamingNear Video on Demand (NVoD)Push ServicesInteractive Gaming


InteractiveInteractive Network ModelNetwork Model

Broadcast Channelunidirectional broadband Broadcast

Channel

Interaction Channelbi-directional Interaction Channel between the service provider and the user

Return Interaction Path (Return Channel) from the user to the service providerForward Interaction Path from the service provider to the user.

Return Channel Satellite Terminal (RCST) formed by

Network Interface Unit (consisting of the Broadcast Interface Module and the Interactive Interface Module)Set Top Unit.

The RCST provides interface for both Broadcast and Interaction Channels


DVBDVB--RCSRCS ArchitectureArchitecture

Network Control Centre (NCC) provides monitoring & control functions

control and timing signals

Traffic Gateway (TG) receives the RCST return signalsprovides accounting functionsinteractive services and/or connections to external publicproprietary and private service providers and network

Feedertransmits the forward link signal

standard satellite digital video broadcast (DVB-S) uplinkare multiplexed the user data and/or the control and timing signals

11


DVBDVB--RCS Burst FormatsRCS Burst Formats

There are four types of bursts:

Traffic (TRF)

ATM (53bytes)

MPEG (188bytes)

Acquisition (ACQ)

Synchronization (SYNC)

Common signaling channel (CSC)

Used by an RCST to identify itself during the log-on.

ACQ and SYNC bursts are required for accurately positioning the RCS Terminal (RCST) burst during and after log-on


Channel Coding and ModulationChannel Coding and Modulation

Coding for channel error protection is applied to traffic and control data

Two alternative coding schemes can be implemented:

Turbo code

Concatenated code

In the case of the concatenated coding, the outer code is a Reed-Solomon

(RS) code and the inner code is a nonsystematic convolutional code

For both coding schemes, a by-passable CRC can be also applied on CSC and SYNC bursts in order to allow error detection

Modulation format is QPSK

EMPS/ASMS 2004 Conference – ESTEC Noordwijk, September 20, 2004

Turbo CodesTurbo Codes

Background:Turbo codes were proposed by Berrou and Glavieux in the 1993 International Conference in Communications.Performance within 0.5 dB of the channel capacity limit for BPSK was demonstrated.

Features of turbo codes:Recursive convolutional encodersParallel code concatenationSerial code concatenationNonuniform or “Pseudo-random” interleavingIterative decoding

12


DVBDVB--RCS Turbo Code OptionRCS Turbo Code Option

Two double-binary Circular Recursive Systematic Convolutional (CRSC) codesThe most significant bit (MSB) of the first byte is assigned to A, the next to B, and so on.Seven coding rates are defined:

1/3, 2/5, 1/2, 2/3, 3/4, 4/5, 6/7These rates are achieved by puncturing the parity bits


Turbo DecoderTurbo Decoder

DEC-1

DEC-2

ΠΠΠΠΠΠΠΠ−−−−1111

sky

pky1

pky2

)(12 ke dL

sky~

)(~

12 ke dL)(21 k

e dL

)(~

21 ke dL

ΠΠΠΠ−−−−1111

)(~

2 kdΛ)(2 kdΛ kd̂


Turbo Code Performance: AWGN ChannelTurbo Code Performance: AWGN Channel

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6

Eb/N0 [dB]

PER

1/3 - QPSK

1/2 - QPSK

2/3 - QPSK

3/4 - QPSK

MPEG packet size

13


DVBDVB--RCS: Multiple AccessRCS: Multiple Access

The multiple-access capability is

either fixed or dynamic slot MF-TDMA

The RCST shall indicate its capability

by using the MF-TDMA field present in

the CSC burst

MF-TDMA allows a group of RCSTs to

communicate with a gateway using a

set of carrier frequencies each of

which is divided into time-slots

The Network Control Center (NCC)

allocates to each active RCST a series

of bursts (slots), each defined by a

frequency, a bandwidth, a start time

and duration


DVBDVB--RCS: more infoRCS: more info

Standard Ref: EN 301 790 Edition: 1.3.1

Interaction for Satellite Distribution Systems

Standard Ref: TR 101 790 Edition: 1.2.1

Guidelines for the Implementation & usage of the DVB Interaction Channel for

Satellite Distribution

www.dvb.org

http://www.dvb.org/documents/white-papers/DVBRCSbkgrbk1sted20021126.pdf


ASMS2004 Tutorials DVB-S2 - ESA ARTES · PDF fileASMS2004 Tutorials DVB-S2 Giovanni E. Corazza...

Documents

Transcript of ASMS2004 Tutorials DVB-S2 - ESA ARTES · PDF fileASMS2004 Tutorials DVB-S2 Giovanni E. Corazza...