Sum-MSE Optimization for Joint Transmission With Limited Backhaul

In this letter, we study beamformer design to minimize the sum mean square error (SMSE) in a multipleinput-multiple-output (MIMO) joint transmission system. By assuming that the backhaul capacity is non-ideal, the original problem is mixed with a l₀ norm, which is non-convex and NP hard. Therefore, we introduce a weighted l₁ norm to relax it and propose a suboptimal iterative scheme to derive the transmit and receive beamformers. This algorithm is proved to be convergent and can be implemented in a distributed way. Simulation shows that our method can achieve an SMSE performance that is similar to the ideal backhaul-based SMSE minimization, whereas the former can decrease 20% backhaul consumption at the same time.