Graph-Based Decoding for High-Dense Vehicular Multiway Multirelay Networks

Densely deployed wireless networks is one of the most important solutions for spectrum shortage expected by 2020 with a huge economic impact. This paper proposes decoding schemes for high- dense vehicular multiway multirelaying (HDV-MWMR) systems comprising multiple multiway relays to serve huge number of users or devices. Due to the nature of huge number, instead of using perfect scheduling, we consider coded random access schemes, where all users transmit their messages uncoordinatedly. Although the transmission is random, the network structures still can be seen as codes-on-the-graph, resembling Low Density Parity Check (LDPC) codes structure, expected to provide an additional gain. The theoretical network capacity bound for HDV-MWMR systems exploiting multiple multiway relays is derived and confirmed via extrinsic information transfer (EXIT) chart analysis and computer simulations. To achieve the network capacity bound, we propose simple decoding schemes based on successive interference cancellation over a sparse graph involving multiple multiway relays. Suitable degree distributions for Rayleigh fading channels are investigated. The results confirm that multiple relays help both on (i) improving the throughput performances, and (ii) capturing the network diversity, which are highly required for future wireless networks.