Linear Transceiver Design for Multi-Carrier Full-Duplex MIMO Decode and Forward Relaying

Abstract

In this paper, we discuss the linear precoding and decoding design problem for a multi-carrier (MC) full-duplex (FD) decode-and-forward (DF) relaying system. We consider the effects of hardware distortions, which contribute to residual self-interference and also inter-carrier leakage. The impact of the imperfect channel estimation is also taken into account. The problem of linear precoding and decoding design is then studied with the goal of maximizing the system sum rate, leading into a non-convex optimization problem. In addition to the traditional per-carrier DF relaying, the case with a joint subcarrier DF is also studied, taking advantage from a group-wise decoding and encoding. An alternating quadratic convex program (AQCP) is then proposed in each case, with a monotonic improvement at each iteration, leading to a guaranteed convergence. Numerical simulations show a significant gain in terms of the end-to-end system sum rate, in particular when the hardware distortions increase and inter-carrier leakage becomes a dominant factor.