Mixed Time Scale Weighted Sum Rate Maximization for Hybrid Beamforming in Multi-Cell MU-MIMO Systems

Abstract

This work deals with hybrid beamforming for the MIMO Interfering Broadcast Channel (IBC), i.e. the Multi-Input Multi-Output (MIMO) Multi-User Multi- Cell downlink. In the case of massive MIMO with a large number of antennas at the base station, it may not be feasible to have as many RF chains ℳ as the number of antennas $N_t$. Hybrid beamforming (BF) involves two stage precoders with an analog stage in RF and a digital stage in baseband. Key assumptions here are that the fully connected analog stage consists of phase shifters and cannot be adapted at the fast fading rate. We consider BF design by maximizing the Weighted Sum Rate (WSR) for the case of Perfect Channel State Information at the Transmitter (CSIT). However, whereas analog and digital beamformers are optimized jointly only from time to time, only the digital beamformers are updated at the fast fading rate with the analog beamformer stage being frozen over the slow fading coherence times. Simulation results show that even if the analog beamformer gets outdated, close to optimal performance gets obtained thanks to the self averaging of large antenna systems. The simulations also show significant improvements compared to suboptimal covariance CSIT based designs for the analog BF.