Predistortion and precoding vector assignment in codebook-based downlink beamforming

This paper considers codebook-based downlink beamforming proposing a linear adaptive channel predistortion method to close the performance gap between codebook and non-codebook based beamforming defined in wireless standards, e.g., LTE-A. The proposed channel predistortion method does not involve any additional signalling overhead. In our novel beamforming concept, we optimize the codebook-based precoding vector assignment, the power allocation, and the channel predistortion matrix jointly to minimize the transmitted power of the base station (BS) while guaranteeing the quality-of-service (QoS) of the mobile stations (MSs) and satisfying the smoothness conditions on the channel predistortion matrix. The joint channel predistortion and codebook-based beamforming (PCB) problem represents a non-convex mixed integer program (MIP). An alternating optimization algorithm (ATOA) and an alternating feasibility search algorithm (AFSA) are developed to approximately solve the PCB problem. Simulation results show that the proposed codebook-based beamforming achieves significant reductions of the transmitted power of the BS and remarkable increases of the percentage of feasible cases of all Monte Carlo runs, as compared to the standard codebook-based beamforming. Numerical results also show that the proposed design performs very close to the non-codebook based beamforming in the given settings and that the proposed ATOA and AFSA are close-to-optimal.