Linear-Equation Ordered-Statistics Decoding

In this paper, we propose a new linear-equation ordered-statistics decoding (LE-OSD). Unlike the OSD, LE-OSD uses high reliable parity bits rather than information to recover codeword estimates, which is equivalent solving system of linear equations (SLE). Only test error patterns (TEPs) that create feasible SLEs, referred as valid TEPs, are used obtain different estimates. We introduce several constraints on Hamming weight TEPs limit overall complexity. Furthermore, analyze block rate (BLER) and computational complexity proposed approach. It shown has similar performance OSD in terms BLER, can asymptotically approach Maximum-likelihood (ML) with proper parameter selections. Simulation results demonstrate significantly reduced compared especially for low-rate codes, usually require order OSD. Nevertheless, reduction also be observed high-rate codes. addition, further improve by applying stopping condition TEP discarding condition. As simulations, improved considerably while maintaining BLER performance, latest from literature.