User Coordination for Fast Beam Training in FDD Multi-User Massive MIMO

Massive multiple-input multiple-output (mMIMO) communications are one of the enabling technologies 5G and beyond networks. While prior work indicates that mMIMO networks employing time division duplexing have a significant capacity growth potential, deploying in frequency (FDD) remains problematic. The two main difficulties FDD scalability downlink reference signals overhead associated with required uplink feedback for channel state information (CSI) acquisition. To address these difficulties, most existing methods utilize assumptions on radio environment such as sparsity or angular reciprocity. In this work, we propose novel cooperative method scalable low-overhead approach to under so-called grid-of-beams architecture. key idea behind our scheme lies exploitation near-common signal propagation paths often found across several mobile users located nearby regions, through coordination mechanism. doing so, leverage recently specified device-to-device capability Specifically, design beam selection algorithms capable striking balance between CSI acquisition multi-user interference mitigation. exploits statistical information, covariance shaping. Simulation results demonstrate effectiveness proposed algorithms, which prove particularly well-suited rapidly-varying channels short coherence time.