Information Set Based Soft-Decoding Algorithm for Block Codes

In this paper we introduce a new sub-optimum soft-decoding algorithm for block codes, based on information sets and erasures, which showed to be very efficient in terms of bit error rate and computational complexity when compared to optimum decoding methods. Our results show that the proposed algorithm performs very close to the maximum likelihood decoder, in terms of bit error rate, with a much lower computational complexity. In addition, we compared the proposed algorithm to a soft-decision Viterbi algorithm for block codes, showing that the proposed method is much less complex than Viterbi. For a given C(n,k) block code, the number of candidate codewords is reduced from 2^n-k to k+1, in the worst case. A set of variants of the proposed IS algorithm is also presented, introducing an interesting trade-off between bit error rate performance and complexity.