Iterative Decoding of Multi-step Majority Logic Decodable Codes
The performance of iterative decoding algorithms for multi-step majority logic decodable (MSMLD) codes of intermediate length is investigated. We introduce a new bit-flipping algorithm that is able to decode these codes nearly as well as a maximum likelihood decoder on the binary symmetric channel. MSMLD codes decoded using bit-flipping algorithms can out-perform comparable BCH codes decoded using standard algebraic decoding algorithms, at least for high bit flip rates (or low and moderate signal to noise ratios).
