In this paper, the decoding algorithm of low-density parity-check (LDPC) codes is analyzed, and a new decoding algorithm based on the belief propagation (BP) algorithm to eliminate the influence of cycles in the factor graph is proposed. In the traditional BP algorithm, the cycles of factor graph sends message back to its source, and this decreases the decoding performance. The new algorithm records each cycle's path and length of each node, and cuts off the path by which message is propagated when the message comes back. It can advance the decoding performance by protect the message of good quality be propagated as widely as possible. Our results of simulation show that the performance of new algorithm is not worse than that of traditional BP algorithm in the low SNR channel and the new algorithm significantly outperform traditional BP algorithm in good channel condition.
[1]
Robert G. Gallager,et al.
Low-density parity-check codes
,
1962,
IRE Trans. Inf. Theory.
[2]
Radford M. Neal,et al.
Near Shannon limit performance of low density parity check codes
,
1996
.
[3]
Brendan J. Frey,et al.
Factor graphs and the sum-product algorithm
,
2001,
IEEE Trans. Inf. Theory.
[4]
David J. C. MacKay,et al.
Good Error-Correcting Codes Based on Very Sparse Matrices
,
1997,
IEEE Trans. Inf. Theory.