Get documents about "The DSP implementation of the decoder - PowerPoint
Document Sample
SpaceOps 2006, 19-23 June, Rome, Italy
Combined Advanced Coding & Modulation
for Future CCSDS High-Rate Missions
Gian Paolo Calzolari, and
Enrico Vassallo
ESA/ESOC
1
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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
2
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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
3
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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
4
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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
5
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
PUNCT.
SYST.
BITS
CC1 FIX PUNCT. CC2
ROW- keep the block length
on the channel constant
INTERLEAVE COLUMN MAPPING/
Rate 1/2 11 Rate 1/2 P/S
R INTERLEAV MODULATION
INPUT 4-states 10 4-states
ER OUTPUT
to 8100 symbols
BITS
SYMBOLS
PUNCT.
PARITY
BITS ATTACHED regardless the
modulation cardinality
FRAME
MARKER
or the code rate.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
6
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
u
c1 inner encoder contains the N+2 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 N+2 parity-check bits
+ D D
generated by the inner encoder. Two
c2 different puncturing algorithms are
+
used to puncture bits.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
7
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
affect the decoding process.
correlation does not adversely MSB
WRITE READ
Row 0
Row 8099
Column 0 Column m-1
Combined Advanced Coding and Modulation
LSB
for Future CCSDS High-Rate Missions
8
SCCC Modulation formats
0111 0011
011 001
0110 0010
010 000
0100 0101 0001 0000
1100 1101 1001 1000
110 100
01 00
8PSK 111 101
16APSK 1110 1010
1111 1011
101100 001100
101110 001110
01111 00111 100110 000110
01101 00101 100100 101101 001101 000100
QPSK 01110 00110 100101
101111
100111
001111
000111 000101
11 10 01100 00100
01010 00010 101001 001001
100001 100011 101011 001011 000011 000001
01000 01001 01011 00011 00000 001010 000010 000000
00001 100000 100010 101010 101000 001000
110000 110010 111010 111000 011000 011010 010010 010000
11000 11001 11011 10011 10001 10000
110001 110011 111011 011011 010011 010001
11010 10010 111001011001
11100 10100
110101 110111 010111 010101
11110 10110
32APSK 110100
111111
111101 011101
011111
010100
11101
11111 10111
10101
64APSK 110110 010110
111110 011110
111100 011100
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
9
Supported set of spectral efficiencies
Nominal values Interleaver constrained
Code
Es/N0 eta K I N m Rate K' I' eta' Es/N0' delta
1 -1.85 0.7254 5,876 8,816 16,200 2 0.36 5,758 8,640 0.7109 -1.96
2 -0.85 0.8659 7,014 10,523 16,200 2 0.43 6,958 10,440 0.8590 -0.89 1.07
QPSK
3 0.15 1.0254 8,306 12,461 16,200 2 0.51 8,398 12,600 1.0368 0.22 1.11
4 1.15 1.2039 9,752 14,630 16,200 2 0.60 9,838 14,760 1.2146 1.21 0.99
5 2.15 1.4012 11,350 17,027 16,200 2 0.70 11,278 16,920 1.3923 2.11 0.90
6 3.15 1.6164 13,092 19,640 16,200 2 0.81 13,198 19,800 1.6294 3.21 1.10
7 2.15 1.4012 11,350 17,027 24,300 3 0.47 11,278 16,920 1.3923 2.11 -1.10
8 3.15 1.6164 13,092 19,640 24,300 3 0.54 13,198 19,800 1.6294 3.21 1.10
8PSK
9 4.15 1.8484 14,972 22,460 24,300 3 0.62 14,878 22,320 1.8368 4.10 0.89
10 5.15 2.0958 16,976 25,466 24,300 3 0.70 17,038 25,560 2.1035 5.18 1.08
11 6.15 2.3568 19,090 28,637 24,300 3 0.79 19,198 28,800 2.3701 6.20 1.02
12 7.15 2.6299 21,302 31,955 24,300 3 0.88 21,358 32,040 2.6368 7.18 0.98
13 6.15 2.3568 19,090 28,637 32,400 4 0.59 19,198 28,800 2.3701 6.20 -0.98
16APSK
14 7.15 2.6299 21,302 31,955 32,400 4 0.66 21,358 32,040 2.6368 7.18 0.98
15 8.15 2.9133 23,598 35,399 32,400 4 0.73 23,518 35,280 2.9035 8.12 0.94
16 9.15 3.2056 25,966 38,951 32,400 4 0.80 25,918 38,880 3.1998 9.13 1.01
17 10.15 3.5053 28,392 42,590 32,400 4 0.88 28,318 42,480 3.4960 10.12 0.99
18 9.15 3.2056 25,966 38,951 40,500 5 0.64 25,918 38,880 3.1998 9.13 -0.99
32APSK
19 10.15 3.5053 28,392 42,590 40,500 5 0.70 28,318 42,480 3.4960 10.12 0.99
20 11.15 3.8111 30,870 46,307 40,500 5 0.76 30,958 46,440 3.8220 11.19 1.07
21 12.15 4.1220 33,388 50,084 40,500 5 0.82 33,358 50,040 4.1183 12.14 0.95
22 13.15 4.4370 35,940 53,912 40,500 5 0.89 35,998 54,000 4.4442 13.17 1.03
23 12.15 4.1220 33,388 50,084 48,600 6 0.69 33,358 50,040 4.1183 12.14 -1.03
64APSK
24 13.15 4.4370 35,940 53,912 48,600 6 0.74 35,998 54,000 4.4442 13.17 1.03
25 14.15 4.7555 38,520 57,782 48,600 6 0.79 38,638 57,960 4.7701 14.20 1.02
26 15.15 5.0766 41,120 61,682 48,600 6 0.85 41,038 61,560 5.0664 15.12 0.92
27 16.15 5.4000 43,740 65,612 48,600 6 0.90 43,678 65,520 5.3923 16.13 1.01
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
10
Simulated performance of the 27
spectral efficiencies (AWGN channel)
1.00E+00
142Mbps 171 207 243 278 326 367 421 474 527 581 640 700 764 824 889 954 1,013 1,078
1.00E-01
1 2 3 4 5 6
1.00E-02
7 8 9 10 11 12
1.00E-03
13 14 15 16 17
BER
1.00E-04
18 19 20 21 22
1.00E-05
23 24 25 26 27
1.00E-06
1.00E-07
1.00E-08
-2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
E s /N 0 [dB]
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
11
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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
12
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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
13
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 >1), 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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
14
Scatter Chart of Parity Check Matrix
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
15
Performances of GSFC Code
The curves
were
determined
at GSFC by
hardware
simulation.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
16
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
k=1024, k=4096, and k=16384.
There are 9 combinations of the 3 block lengths with
the 3 possible code rates providing flexible solutions to
different mission needs.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
17
Constant Frame Length
By choosing to keep the information block length “k”
constant among family members, rather than the
codeblock length “n”, the spacecraft’s 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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
18
The JPL Parity Check Matrix
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
19
Performances of JPL Code
From left to right
the performance
curves for the
midsize information
block length codes
with parameters
(n=8192, k=4096)
rate 1/2, (n=6144,
k=4096) rate 2/3,
and (n=5120,
k=4096) rate 4/5.
The curves were
determined at JPL
by hardware
simulation.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
20
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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
21
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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
22
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.
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
23
Thank you for your attention.
Questions ?
Combined Advanced Coding and Modulation
for Future CCSDS High-Rate Missions
24
