Extrinsic Information Transfer Analysis and Design of Block-Based Intermediate Codes

Download

PDF - Published Version Download (315Kb)

Description/Abstract

Intermediate codes have been shown to facilitate iterative decoding convergence to the maximum likelihood error ratio performance in serially concatenated schemes. In this paper, we propose a novel block-based intermediate code as an alternative to classic convolutional intermediate codes. Since it is block-based, our intermediate code facilitates practical implementations having reduced memory and processing requirements. Furthermore, we demonstrate that it is simpler to analyse and optimise the iterative decoding process, when our block-based intermediate code is employed instead of a convolutional intermediate code. Finally, we demonstrate that the proposed block-based intermediate code facilitates significantly reduced error ratios in practical schemes, when employing short transmission frames and a limited decoding complexity.