Diversity-Multiplexing Tradeoff Analysis of Coded Multi-User Relay Networks

We study the impact of multiple access strategies on the diversity-multiplexing tradeoff (DMT) performance in wireless multi-user relay networks. The networks contain multiple independent sources, multiple half-duplex decode-and-forward (DF) relays, and one common destination. Instead of separately retransmitting each source message, the relays employ a class of spectrally efficient finite field network codes to assist the sources. It is shown that fully orthogonal or fully non-orthogonal transmission among sources/relays does not necessarily provide optimized DMT performance. We propose a novel transmission protocol that divides the sources and relays into individual clusters. The nodes within one cluster transmit non-orthogonally while the transmissions of different clusters span orthogonal channels. We provide the method to calculate the achievable DMT for each clustering strategy. The network DMT performance can thus be optimized by properly clustering the multiple sources and relays.