Hybrid Precoding Using Out-of-Band Spatial Information for Multi-User Multi-RF-Chain Millimeter Wave Systems.

With a tremendous amount of underutilized bandwidth, millimeter wave (mmWave) communication is a promising solution to meet the increasing data-rate demands. To deal with high pathloss in mmWave frequency band, efficient large antenna arrays become crucial to provide necessary beamforming and spatial multiplexing gains. Hybrid analog/digital beamforming has been an encouraging technique in order to reduce the complexity and power consumption of large antenna arrays. In this paper, we propose a low-complexity two-stage hybrid analog/digital precoding and combining scheme for downlink multi-user multi-RF-chain mmWave systems using out-of-band spatial information to minimize the mean-squared error (MMSE) in receiving data streams. Particularly, according to sparse spatial characteristics of the mmWave channel, we utilize more efficient Grassmannian codebooks as the training codebooks to generalize a strategy of utilizing sub-6GHz spatial information for constructing multiuser mmWave communication links. The multiuser analog beamforming problem is then formulated as a weighted sparse signal recovery problem with the weights obtained from sub-6GHz spatial information. In addition, the digital precoding and combining scheme is derived using an efficient MMSE scheme for multi precoders at transmitters and combiners at receivers. Numerical results show the obvious advantages of the proposed multi-user hybrid analog/digital precoding/combining design for various practical scenarios.