On Combining Distributed Joint Source-Channel-Network Coding and Turbo Equalization in Multiple Access Relay Networks

We consider a multiple access relay channel where two sources S1 and S2 send their data to a common destination node aided by a shared relay node. In this paper we extend previous work on this topic by assuming that the output of both information sources are correlated with memory. Furthermore, we consider frequency-selective fading channels, which introduce inter-symbol interference among the transmitted symbols. In this context, we propose an iterative receiver which comprises Turbo Maximum a Posteriori (MAP) equalization and distributed source-channel-network decoding in an iterative fashion. Various relaying strategies such as network coding or time multiplexing are analyzed in detail. In addition, a suboptimal but less complex Minimum Mean Square Error (MMSE) linear equalizer is also utilized to account for communication scenarios with constrained complexity. Simulation results show that network coding based strategy outperforms, in all cases, the other considered forwarding schemes.