Abstract
This paper deals with bit error performance analysis of network-coded (NC) multiple access relay channels that operate in the decode and forward mode with error-prone relaying under quasi-static Rayleigh fading channels. It shows that the combination of channel coding and NC relaying can be regarded as a distributed coding scheme (DCS) if the error propagation problem is adequately addressed. For this purpose, methods based on NC soft information relaying (NC-SIR) and NC hard information relaying (NC-HIR) are investigated. In the framework of NC-SIR, the recently proposed Rayleigh-Gaussian log-likelihood-ratio-based Model is used for modelling the soft estimated symbols at the output of the relay soft encoder. The resulting scheme is referred to as NC-SIR-DCS. On the other hand, two NC-HIR-based schemes are studied and analyzed. The first employs automatic repeat request (ARQ) protocols between the relay node and each source node to ensure error-free propagation and is referred to as NC-ARQ-DCS, while the second performs a limiter function at the destination during iterative decoding as introduced in Thobaben (1) distributed and is referred to as NC-LFD-DCS. The SNR of the one-hop link, the equivalent to the sources-relay-destination links, is estimated in the case of NC-SIR-DCS and is lower bounded in the case of NC-LFD-DCS, while in the case of NC-ARQ-DCS, the average SNR cost per information bit due to the ARQ protocol is estimated. These processes allow tractable analysis of the bit error rate performances of the presented network-coded DCSs. Simulation results are carried out to assess the accuracy of the proposed bounds for different relay positions between source nodes and destination.