Page 364 - Satellite Communications, Fourth Edition

344 Chapter Eleven

Code rate vs Performance
E /N at BER = 10 −8
0
b
1.0
0.9
0.8

Code rate 0.7
0.6
Shannon limit (BPSK)
0.5
TPC 4K blocks
0.4 TPC 16K blocks
Reed-Solorn on/Vlterbi
0.3 LDPC 30K
LDPC 16K
0.2
0.0 1.0 2.0 3.0 4.0 5.0
E /N (dB)
b
0
Figure 11.14 FEC performance of Comtech AHA 4701 LDPC coder. (Courtesy of A.
Summers Comtech AHA.)
be noticed that the estimate from any given parity node is not included
in the new estimate sent to that parity node. This process is repeated
a number of times (iterated) before a final “hard decision” is made.
Turbo codes and LDPC codes are being employed in a number of satel-
lite systems with a resulting improvement in channel performance, for
example the Digital Video Broadcast S2 standard (DVB-S2) employs
LDPC as the inner code in its FEC arrangement (Breynaert, 2005,
Yoshida, 2003), and the DVB-RCS plans to use turbo codes (talk Satellite,
2004). The performance of a number of codes is shown in Fig. 11.14.

11.12 Automatic Repeat Request (ARQ)
Error detection without correction is more efficient than FEC in terms
of code utilization. It is less complicated to implement, and more errors
can be detected than corrected. Of course, it then becomes necessary for
the receiver to request a retransmission when an error has been
detected. This is an automatic procedure, termed automatic repeat
request (ARQ). The request for retransmission must be made over a
low-bit-rate channel where the probability of bit error can be kept neg-
ligibly small. Because of the long round-trip delay time, on the order of
half a second or more, encountered with geostationary satellites, ARQ
is only suitable for transmission that is not sensitive to long delays. ARQ
is normally used with block encoding.
An estimate of the probability of an error remaining undetected can
k
be made. With an (n, k) block code, the number of datawords is 2 ,and