In this work, we study the linear precoding and decoding design problem for a full-duplex (FD) multi-carrier (MC) decode and forward (DF) relaying system, under multiple sources of impairments. In particular, the impact of non-linear hardware distortions at the transmit and receiver chains, leading to residual self-interference (SI) and inter-carrier leakage (ICL), as well as the imperfect channel state information (CSI) are taken into account. In the first step, the known time-domain characterization of hardware impairments is transformed over a general orthonormal MC basis. As a result, the problem of linear precoding and decoding design is formulated to maximize the MC system sum-rate, however, leading to a non-convex mathematical structure. An alternating quadratic convex program (AQCP) is consequently proposed, with a monotonic improvement at each iteration, leading to a guaranteed convergence. The proposed AQCP framework is then extended, employing the Dinkelbach algorithm, in order to maximize the system energy efficiency in terms of bits-per-Joule. The proposed design strategies are considered both for per-subcarrier as well as the joint-subcarrier coding strategies, wherein the latter case the coding is performed jointly over all subcarriers. Numerical simulations show a significant gain in the performance of the proposed algorithms compared to the half-duplex (HD) counterparts or to the solutions where the impact of impairments are not considered, particularly when the hardware accuracy is not very high.
Read full abstract