Summary
Asia-Pacific Conference on Communications
2008
Session Number:16-AM2-B
Session:
Number:1569126981
An Efficient BF LDPC Decoding Algorithm Based on A Syndrome Vote Scheme
Jui-Hui Hung, Sau-Gee Chen,
pp.-
Publication Date:2008/10/14
Online ISSN:2188-5079
DOI:10.34385/proc.27.1569126981
Summary:
This work presents a high-performance bit-flipping (BF) algorithm utilizing a proposed syndrome vote technique of a defined culprit bit set, named CVBF algorithm for LDPC decoding. It can achieve significant decoding performance improvements by about 0.9dB over the most efficient BF algorithm, owing to the introduced an additional syndrome vote procedure after updating the flipping reliabilities of all the received bits in each iteration. Moreover, its performance is comparable to the min-sum algorithm (MSA) under the condition of same iteration number, but with much lower complexity per iteration. The proposed syndrome vote scheme only costs little overhead in hardware realization. Besides, an early termination strategy suited for the proposed algorithm is also devised to further reduce the iteration number.