Pragmatic approach to soft-decision decoding of linear block codes

The complexity of algorithms to perform soft-decision decoding on block codes has impeded their use in practical systems. The authors investigate correlation decoding along with hard-decision decoding. The proposed method uses a hard-decision decoder that decodes to a codeword. Although this codeword may be in error, it is usually close in Hamming distance to the transmitted codeword. Correlation is then performed on only a small set of `nearby´ codewords. This pseudo maximum likelihood (PML) decoding method can approach the performance of true correlation decoding with a significant reduction in complexity