Iterative Decoding of Block Turbo Codes over the Binary Erasure Channel

This paper investigates iterative decoding of product codes over the binary erasure channel. A simplified row-column iterative decoding algorithm combining bit guessing with bounded-distance erasure decoding is introduced. This algorithm shares many similarities with the original Chase-Pyndiah turbo decoding algorithm and offers a wide range of trade-offs between performance and complexity. The performance of the proposed iterative decoding algorithm has been evaluated by Monte-Carlo simulation for several selected high-rate product codes. Both binary Bose-Chaudhuri-Hocquenghem and non-binary Reed-Solomon component codes have been considered. The results show that product codes provide near-capacity performance without exhibiting error floors in the simulated word-erasure-rate region.