Optimal Beamforming in MIMO Two-Way Relay Channels

In this paper, the optimal analog network coding strategy (or relay beamforming) in MIMO two-way relay channels is studied. The weighted sum rate is used as the optimization criterion. Motivated by the classic Arimoto-Blahut algorithm, an iterative algorithm, which monotonically improves the cost function, is designed to update the relay beamformer. This relay beamforming algorithm is further extended to the case of joint source and relay beamforming. We show that in the iteration of the algorithms, the updates of the relay beamforming matrix and the source beamforming matrices are accomplished by solving a series of quadratic programs and quadratically constrained quadratic programs, which are of low complexity. Simulations are presented to compare the proposed algorithms with the performance upper bounds and the existing algorithms.