1: \begin{abstract}

2: %Hybrid analog and digital beamforming (HBF) has great potential for providing spatial multiple access in multiuser millimeter wave (mmWave) communication systems at affordable cost and power consumption. However, the HBF optimization is more challenging than that in the point-to-point case as the signals at different users cannot be cooperatively processed. In this paper, we investigate the HBF based on the minimum mean square error criterion. Using the alternating minimization principle, the original HBF problem is decoupled into the base station precoding and users combining sub-problems. We first derive the optimal digital precoding matrix and optimize the analog one via general eigen-decomposition (GEVD). With the help of the unitary constraint on the digital combiners, we then derive a closed-form expression of the digital combiners, and show that the analog ones can also be optimized via GEVD. Simulation results show that the proposed MMSE HBF scheme is able to approach the performance of the fully digital one.

3: This letter investigates the hybrid analog and digital beamforming (HBF) design for multiuser millimeter wave (mmWave) communication systems based on the minimum mean square error (MMSE) criterion. Using the alternating minimization method, the hybrid precoder of the base station (BS) and the hybrid combiners of the users are alternatively optimized. It is shown that both the optimized digital precoder of the BS and the digital combiners of the users have closed-form expressions, and their corresponding analog ones can be efficiently obtained via generalized eigen-decomposition. Simulation results show that the proposed MMSE HBF scheme has fast convergence and performs close to the fully digital beamforming.

4: \end{abstract}

5: