©2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

Recently, a receive spatial modulation (RSM) for massive multiple-input-multiple-output operating in millimeter wave (mmWave) was introduced with the purpose of simplifying user terminal circuit by employing only one radio-frequency chain and attaining high spectral efficiency by exploiting the receive spatial dimension. However, when RSM is applied in a mmWave channel, it demands a challenging receive antenna selection (RAS) procedure. On the other hand, the power consumption at the transmitter side is high when a full digital (FD) precoder is envisioned. We consider the joint problem of RAS and precoder designs based low complexity hybrid architecture. For the sake of simplicity, we divide this problem into two subproblems. First, we design the RAS assuming FD precoder, and then, we design the hybrid precoder. We propose two novel and efficient RAS methods. First, we formulate the RAS as non-convex optimization problem. Then, we convert it into a convex optimization problem by introducing novel lower bounds and relaxing non-convex constraints. Second, we provide sequential algorithms that approach the optimal selection where we (add/remove) one (good/poor) antenna per iteration. We propose novel zero forcing hybrid precoder based convex optimization that maximizes the received power. We prove that the proposed precoder is optimal when the channel is highly spatially sparse. The proposed designs have been compared with the best known methods in terms of average mutual information and energy efficiency showing significant improvements.

Peer Reviewed