A comparison of SISO algorithms for iterative decoding of multidimensional product codes

In this paper, we consider iterative decoding of multidimensional product codes (MPC). Such codes, also called iterated codes, were first introduced by Elias (1954). After a formal definition of MPG, we compute their low-weight distribution. Using this distribution, we derive an improved asymptotic union bound on maximum likelihood (ML) decoding of such codes over a Gaussian channel. The main part of the paper is devoted to a comparison of efficient SISO algorithms, used as basic tools in the iterative decoding process. Simulation results show that MPC may be more suitable than any other schemes for packet wireless transmission involving packet sizes of less than 1000 information bits and requiring very low packet error rates (PER).