A Decomposition Approach to MIMO Interference Relay Networks

In this paper, we consider a class of wireless ad hoc networks, namely two-hop MIMO interference relay network, where L M-antenna source-destination pairs communicate via K relay nodes each equipped with N antennas. Based on recent advances on characterizing the limits of MIMO X channel and the concept of interference alignment, we treat this network as the cascade of two MIMO X networks and decompose it into L non- interfering parallel MIMO relay channels. The rationale behind this approach can be justified by the information theory of MIMO X networks which reveals that they provide higher degrees of freedom than the conventional MIMO interference channels. Signaling, interference alignment, relay selection and simulation results will be presented. Compared to the well-known distributed interference cancellation technique, this approach offers higher capacity for small to medium number of relay nodes at the cost of feeding back CSI from the relays to the source nodes and more advanced signaling.