Relay selection and residual self-interference mitigation for cognitive full-duplex two-way MIMO relaying system

In this paper, we consider the problem of relay selection in a cognitive full-duplex two-way multiple-input multiple-output relaying network. The cognitive nodes operate on the same frequency as the licensed primary user (PU). The end-to-end communication between the secondary users (SU) via a secondary relay (SR) takes two time slots. The SR operates on amplify-and-forward (AF) protocol. In the first time slot, the SR relay with the highest receive signal-to-noise ratio is selected as the optimal relay. At the same time, the precoders at the SU are designed to limit the interference to the PU below a certain threshold. All the nodes in the network are assumed to have imperfect knowledge of channel state information of their loopback self-interference channels and hence experience residual self-interference (RSI). After selecting the optimal SR, in the next time slot, we design the SR precoder and SU receive filters, to mitigate the RSI. We do so by minimizing the sum of mean square error of the end-to-end communication with a transmit power constraint at SR. To curb the effect of cumulative RSI caused by the AF operation of the SR, we update the SR precoder and SU receive filters at each time slot. The efficacy of our proposed designs is illustrated in the simulation results.