A novel beamforming method for multiple peer-to-peer relay networks

One multi-antenna relay can support multiple peer-to-peer communications in AF (amplify-and-forward) manner. All sources transmit their signals to the relay in the first time slot. In the second time slot, the relay multiplies its received signal with an amplifying matrix, and broadcasts it to the destinations. We propose a novel method to specify the amplifying matrix in which we decompose the amplifying matrix into two parts. The first part is an MMSE (minimum mean square error) receiving matrix, and the second part is a beamforming matrix based on the maximization of VSLNR (virtual signal-to-leakage-plus-noise ratio). To apply our proposed algorithm, we first propose two simple power allocation methods: EPA (equal power allocation) and EGA (equal gain allocation), and then propose an optimization method using projection gradient algorithm. Numerical results show that our proposed algorithm outperforms the existing ZFBF (zero-forcing beamforming) algorithm and the optimal power allocation improves system performance significantly under asymmetric channel conditions.