Hybrid Beamforming Optimization for DOA Estimation Based on the CRB Analysis

Direction-of-arrival (DOA) estimation is one of the most demanding tasks for millimeter wave (mmWave) communication massive multiple-input multiple-output (MIMO) systems with hybrid beamforming (HBF) architecture. In this paper, we focus on optimization HBF matrix receiving pilots to enhance DOA performance. Motivated by fact that many existing algorithms can achieve Cram\'{e}r-Rao bound (CRB), formulate problem aiming at minimizing CRB prior knowledge rough range. Then, tackle intractable non-convex constraint introduced analog beamformers, propose an efficient manifold (MO) based algorithm. Simulation results demonstrate significant improvement proposed CRB-MO algorithm over conventional random algorithm, and provide insights design in beam training stage practical applications.