Comparison of decoding turbo Gallager codes in hybrid decoding arrangements with different iterative decoders over the erasure channel

In this paper, different iterative decoders for turbo Gallager codes are optimised and compared for the binary erasure channel. The complexity and performance differences between turbo decoder, BCJR-based Look-Up Table decoder and belief propagation decoder are analysed and evaluated. A hybrid decoding arrangement, which uses an iterative decoder followed by a maximum likelihood ¿In-Place¿ matrix inversion algorithm, is compared for the different iterative decoders. Results are presented which show that the BCJR-based iterative decoders achieve better performance than using the belief propagation decoder for turbo Gallager codes in the erasure channel. When small encoder memory is selected, the optimised Look-Up Table decoder provides a good balance between convergence performance and complexity.