EXIT Chart Based Joint Source-Channel Coding for Binary Markov Sources

Abstract

In this paper, we propose a new joint source-channel decoding technique for transmitting binary Markov sources over AWGN channels. Our approach is based on serially concatenated coding and code doping for inner code. By combining the Markov source and the outer code trellis diagrams, a super trellis is constructed to exploit the time-domain correlation of the source. A modified version of BCJR algorithm is derived based on this super trellis, that can achieve considerable gain in terms of mutual information. The standard BCJR algorithm is used for decoding of inner code where code doping is adopted for better matching of extrinsic information transfer (EXIT) characteristics. EXIT chart analysis is performed to investigate convergence property of the proposed technique and to optimize the code parameters. Simulation results for bit error rate (BER) evaluation and EXIT chart analysis indicate that the proposed technique can achieve significant gains over the system in which source redundancy are not exploited, and thereby the BER performance of the proposed system is very close to the Shannon limit.