Joint Decoding: Extracting the Correlation among User Pairs in a Multi-way Relay Channel

Abstract

This paper describes a novel mechanism for joint decoding of the network coded symbols in a multi-way relay node. The mechanism, based on belief propagation algorithm, utilizes the correlation between adjacent network coded symbols to minimize the error propagation problem significantly, compared with previous methods. In case of increasing degree of asynchrony, disjoint decoding exhibits poorer error performance, whereas joint decoding helps to maintain the performance level close to that in the synchronous case both in additive white Gaussian noise and fading channels. Thus, this method adds robustness to the multi-way relay channel against channel imperfections like asynchronism and fading in practical propagation environments.