BWT based universal lossless source controlled channel decoding with low density parity check codes

Summary form only given. In many channel decoding applications, redundancy is left in the channel coded data. A new method for utilizing this redundancy in channel decoding is proposed. The method is based on the Burrows-Wheeler transform (BWT) and on universal compression techniques for piecewise stationary memoryless sources (PSMS), and is applied to regular low-density parity-check (LDPC) codes. Two settings are proposed. In the first, the BWT-PSMS loop is in the decoder, while in the second, the rearrangement of the data is performed with the BWT before channel encoding, and then the decoder is designed for extracting statistics in a PSMS. After the last iteration, the data is reassembled with the inverse BWT. Simulations show that the bit error rate performance of the new method (in either setting) is almost as good as genie-aided decoding with perfect knowledge of the statistics.