Combined Advanced Coding

1
Combined Advanced Coding Modulation for Future
CCSDS High-Rate Missions
Gian Paolo Calzolari, and Enrico
Vassallo ESA/ESOC
2
Introduction

• TRADITIONALLY in CCSDS coding and modulation
  techniques have been kept separated from each
  other and assigned to two different WGs.
• Mission designers are (normally) free to select
  whatever CCSDS channel code and couple it with
  any of the CCSDS modulations.
• This approach is sub-optimal. Already in 1999,
  CCSDS recognized the advantages of combining
  modulation and coding and introduced 'Trellis
  coded modulation' for a specific class of
  missions
• As more and more missions share the scarce
  spectrum resource with continuously increasing
  data rate requirements while the on-board power
  remains constrained, it is necessary to extend
  such techniques to the other frequency bands used
  by the space science services.

3
The 3 Approaches (1 of 2)

• The first option proposed by the French Agency
  CNES is to adopt as is the recently developed
  DVB-S2 standard for the next generation of
  digital broadcast via satellite.
• The second option by ESA is based on a serial
  concatenation of a Turbo-like code (SCCC) coupled
  with QPSK, 8PSK, 16APSK and even higher order
  modulations.
• Such scheme was proposed by ESA to the DVB-S2
  forum in 2004 where it was the runner-up due to
  the complexity of the decoding relative to the
  selected LDPC while achieving basically the same
  power efficiency. Further work by ESA in 2005 led
  to the discovery of a parallelization method
  making the decoding simpler than DVB-S2 LDPC by
  at least 30, thereby promising higher throughput
  for the same complexity or lower complexity for
  the same data rate.
• Both CNES and ESA approaches require several
  changes to the various layers of CCSDS.

4
The 3 Approaches (2 of 2)

• The third scheme proposed by NASA is based on the
  LDPC codes being considered by CCSDS channel
  coding WG, pragmatically coupled with QPSK
  modulation or higher order schemes like 8PSK.
• Being a pragmatic approach based on existing
  CCSDS standards or proposed standards, it is
  designed to fit CCSDS layers seamlessly at the
  possible expenses of performances.
• Actually NASA has provided two flavors with two
  kinds of LPDC codes based on different design
  approaches.
• The difficulty in finding consensus has led to
  explore the new approach of producing Agency
  specific experimental Orange Books that may
  eventually allow standardization on a faster
  track.
• In addition, it is planned to increase the
  future work program for the joint effort by
  coding and modulation delegates in this area.

5
CNES Proposal for DVB-S2 Adoption

• In spring 2002 the Digital Video Broadcasting
  (DVB) Project initiated the search for a second
  generation standard for broadband satellite
  applications DVB-S2.
• At the end of 2002 a LDPC code based solution
  proposed by Hughes Network Systems was selected
  and eventually converted into the European
  Standard (Telecommunications series) ETSI EN 302
  307. Such scheme is based on a pragmatic coupling
  of LDPC codes with QPSK and 8PSK modulations
• It is expected that hardware for space science
  applications may become available as spin-off of
  the DVB-S2 market although the cost of the
  applicable patents would have to be taken into
  account.
• However this standard is not directly applicable
  to CCSDS data structures and CNES, to support
  their proposal, investigated the main features
  and performance of the channel coding scheme
  selected by DVB-S2 with respect to frames shorter
  that those under definition by DVB. In their
  presentation at CCSDS, CNES showed the
  possibility of using frame sizes compatible with
  CCSDS at expenses of performance reduction with
  respect to DVB-S2 frames.

6
The SCCC scheme proposed by ESA

• It is based on the serial concatenation of an
  outer 4-state systematic recursive rate ½ encoder
  punctured to rate 2/3, an interleaver and an
  inner 4-state systematic recursive rate ½ encoder
  with suitable puncturing to obtain the desired
  rate
• The code design involves choosing the puncturing
  patterns matching the desired rate

The interleaver length is designed in order to
keep the block length on the channel constant to
8100 symbols regardless the modulation
cardinality or the code rate.
7
Convolutional Encoders

• The SCCC is based on the serial concatenation of
  two identical 4-state systematic recursive rate ½
  encoders. The outer convolutional encoder is
  punctured through a fixed scheme to a rate 2/3
• The outer convolutional encoder is punctured
  through a fixed scheme to a rate 2/3
• In order to obtain the desired coding rate,
  puncturing is performed at the output of the
  inner encoder.

The upper register at the output of the inner
encoder contains the N2 inner systematic bits,
which coincide with the interleaved outer code
word plus the 2 bits terminating the inner
trellis. The lower register, instead, contains
the N2 parity-check bits generated by the inner
encoder. Two different puncturing algorithms are
used to puncture bits.
8
Row-column interleaver

• At the transmitter side a row-column interleaving
  is used to spread the bits belonging to one
  symbol (pragmatic approach). The interleaving
  size is equal to the size of one codeword and the
  number of columns is equal to m, where m is the
  efficiency of the modulation scheme.
• The bit-interleaving is such that the bits
  transmitted with the same modulation signal are
  spread at the output of the inner encoder so that
  their correlation does not adversely affect the
  decoding process.

9
SCCC Modulation formats
8PSK
16APSK
QPSK
32APSK
64APSK
10
Supported set of spectral efficiencies
11
Simulated performance of the 27 spectral
efficiencies (AWGN channel)
12
NASA Proposal for LDPCC

• NASA found the answer to bandwidth efficient
  codes in Low Density Parity Check Codes (LDPCC).
• Opposite to the LDPC Codes selected for DVB-S2,
  the NASA proposals consider codes designed to fit
  with traditional CCSDS data structures.
• Researches performed at the Goddard Space Flight
  Center (GSFC) in Maryland and at the Jet
  Propulsion Laboratories (JPL) in California have
  actually ended up into two different flavors of
  this approach.

13
Low Density Parity Check Code for Rate 7/8

• The proposal by NASA/GSFC has started in fall
  2002 with the submission to CCSDS Channel Coding
  Panel 1B of a White Paper based on Euclidean
  Geometry LDPCC.
• The rationale was that this type of codes had
  shown to provide very low error floors and very
  fast iterative convergence, important qualities
  for near Earth applications where very high data
  rates and high reliability are the driving
  requirements.
• The LDPC code considered by NASA/GSFC is a member
  of a class of codes called Quasi-Cyclic codes.

14
Construction of GSFC Codes

• The construction of these codes involves
  juxtaposing smaller circulants (or cyclic
  submatrices) to form a larger parity check or
  base matrix.
• Being a Circulant a square binary matrix where
  each row is a cyclic shift of the row above
  (degree may be gt1), the GSFC matrix of circulant
  is built such that every row is one bit right
  cyclic shift (where the end bit is wrapped around
  to the beginning bit) of the previous row.
• Constructing parity check matrices in this manner
  produces two positive features 1. the encoding
  complexity can be made linear with the code
  length or parity bits using shift registers, and
  2. encoder and decoder routing complexity in the
  interconnections of integrated circuits is
  reduced.
• With this approach a baseline (8176, 7156) LDPC
  code has been designed. The rate of this code is
  (7156/8176 0.875 i.e. approximately 7/8). A
  total of 7156 information bits are used (894.5
  octets).
• The parity check matrix for this code is formed
  by using a 2 x 16 array of 511 x 511 square
  circulants creating a parity check matrix of
  dimension 1022 x 8176. A scatter chart of the
  parity check matrix for the rate 7/8 LDPC code is
  shown in next slide where every 1 bit in the
  matrix is represented by a point.

15
Scatter Chart of Parity Check Matrix
16
Performances of GSFC Code

• The curves were determined at GSFC by hardware
  simulation.

17
Low Density Parity Check Code Family

• The proposal by NASA/JPL includes a complete
  family of LDPC Codes identified according to
  well defined criteria.
• The selected code rates are 1/2, 2/3, and 4/5.
  These values are about uniformly spaced by 1 dB
  on the rate-dependent capacity curve for the
  binary-input AWGN channel. The selected
  (information) block lengths are k1024, k4096,
  and k16384.
• There are 9 combinations of the 3 block lengths
  with the 3 possible code rates providing flexible
  solutions to different mission needs.

18
Constant Frame Length

• By choosing to keep the information block length
  k constant among family members, rather than
  the codeblock length n, the spacecrafts
  command and data handling system can generate
  data frames without knowledge of the code rate.
• The selected codes are systematic.
• They are based on Accumulate Repeat Accumulate
  Codes, precisely Accumulate Repeat-4 Accumulate
  (AR4) codes.

19
The JPL Parity Check Matrix
20
Performances of JPL Code

• From left to right the performance curves for the
  midsize information block length codes with
  parameters (n8192, k4096) rate 1/2, (n6144,
  k4096) rate 2/3, and (n5120, k4096) rate 4/5.
• The curves were determined at JPL by hardware
  simulation.

21
Orange Books

• Such a wide offer of choices did not make easy to
  get consensus on selecting a single proposal. The
  difficulty in finding consensus has led to
  explore the new approach of producing Agency
  specific experimental Orange Books, i.e.
  Experimental Specifications.
• The "Experimental" designation typically denotes
  a specification that is part of some research or
  development effort. Its funding and other
  associated resources are normally independently
  provided by the organization that initiates the
  work.
• This designation therefore allows the work to
  progress roughly to the equivalent technical
  status of a Draft Standard without being
  actually on the Standards Track.
• Experimental work may be rapidly transferred onto
  the Standards Track if a hard requirement
  emerges, thus shortening the response time in
  satisfying the new customer.

22
Why Orange Books?

• Opposite to other fields where the impact of
  hardware solutions is less relevant, in Channel
  Coding as well as in RF and Modulation it is more
  difficult to reach consensus via compromises
  merging features taken from more proposals.
• Orange Books have the advantage that a period of
  time may elapse allowing to reconsider the
  available solutions at the light of flying space
  mission and progress.
• Orange Books shall not be seen as way to escape
  discussion and aim for consensus, but as an
  effort to record important work for future
  verification and re-discussion according to
  progress in requirements and technology.

23
Conclusion

• The very lively discussion within CCSDS Working
  Groups has confirmed that the achievement of
  gains in both the spectral and the power
  efficiency domains over conventional CCSDS
  encoding followed by binary and quaternary (BPSK
  and QPSK) PSK modulations is essential.
• Therefore, the planning of work to better
  investigate the applicable requirements, possibly
  narrowing the mission scenarios, is a key point
  for the future together with the inclusion of the
  modulation aspects to complement coding
  techniques.
• For these reasons, it is planned to increase the
  future CCSDS work program for the joint effort by
  coding and modulation delegates in this area.

24
Thank you for your attention.

• Questions ?