Design optimization of low density parity check codes in multipath underwater acoustic channels

Abstract

We propose an iterative decoding and equalizing scheme using low density parity check codes and decision feedback equalizer for underwater acoustic communication system. The scheme decodes and equalizes iteratively by linking a LDPC decoder with an adaptive DFE and feeding output soft information of the LDPC decoder to the DFE as priori information. We extend extrinsic information transfer (EXIT) charts to analyze the performance of LDPC codes in multipath underwater acoustic channels. Furthermore, we introduce an EXIT aided method of optimally designing the parameters of LDPC codes using modified differential evolution algorithm in multipath underwater acoustic channel. Design example and simulation are presented in two different underwater acoustic channels which are generated by BELLOP ray tracing model. Bit error rate of the proposed scheme is less than 10-5 as signal to noise rate (SNR) is greater than 8.5 dB in the channels. Compared the regular LDPC code with degree 3 and the Turbo code with feedback and feedforward generator polynomials 1+D+D 2+D 3 and 1+D 2+D 3, the optimized LDPC codes achieve gains of about 0.5 dB and 0.8 dB, respectively, at BER= 10-4 under the same code rate and block length.