LDPCA Code Construction for Slepian-Wolf Coding

摘要

Error correcting codes used for Distributed Source Coding (DSC) generally assume a random distribution of errors. However, in certain DSC applications, prediction of the error distribution is possible and thus this assumption fails, resulting in a sub-optimal performance. This letter considers the construction of rate-adaptive Low-Density Parity-Check (LDPC) codes where the edges of the variable nodes receiving unreliable information are distributed evenly among all the check nodes. Simulation results show that the proposed codes can reduce the gap to the theoretical bounds by up to 56% compared to traditional codes.