A robust hybrid beamforming algorithm for millimeter-wave (mmWave) massive multiple-input multiple-output (MIMO) systems with imperfect channel state information (CSI) is presented. The imperfect CSI is modeled as a norm ball/bounded model assuming that the actual channel lies within the neighborhood of the estimated channel, leading to the worst-case robust design. In particular, we first utilize the S-procedure to transform the non-convex problem with uncertain CSI into an easily solvable convex semidefinite programming (SDP) problem, thereby obtaining the optimal unconstrained beamformer. Then, we derive the hybrid beamformer that approximates the optimal one by alternating optimization. In this step, the non-convex problem arising from the constant modulus constraints is addressed by the successive convex approximation (SCA). Numerical simulations validate the theoretical analysis and demonstrate that the proposed algorithm is robust to imperfect CSI.
Robust Hybrid Beamforming Design for mmWave Massive MIMO Systems with Imperfect CSI
24.06.2024
441576 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch