Joint Symbol-Level Precoding and Radiation Pattern Design for Downlink Reconfigurable MIMO

Pattern reconfigurable multiple-input multiple-output (PR-MIMO) can manipulate the wireless channel according to different communication requirements. In this paper, we discuss the potential of constructive interference (CI)-based symbol-level precoding (CI-SLP) in PR-MIMO communication systems. The joint design problem that optimizes the SLP strategy and the radiation pattern of PR-MIMO for phase-shift keying (PSK) modulation is formulated to maximize the worst serviced user’s communication quality. Since the optimization variables are softly-coupled, we employ the alternating optimization framework to decompose the joint design problem into the SLP design sub-problem and the pattern design sub-problem. We simplify the pattern design sub-problem and propose an interior-point algorithm, where a sequential optimization-based scheme is further proposed as a sub-optimal solution with low complexity. Furthermore, the discussion is extended to quadrature amplitude modulation (QAM) modulated systems, where a special stopping criterion is proposed to guarantee the performance gain of the proposed scheme. The practical realization of the designed reconfigurable antenna array is also discussed, where we propose a design scheme using programmable metasurface antennas based on time-division switching to enable quick and adaptive pattern reconfiguration. Numerical results demonstrate that the radiation pattern configurability can further enhance the benefit of SLP over conventional precoding approaches.

I. Introduction

Compared with the single-antenna technique, the superiority of Multiple-Input Multiple-Output (MIMO) technology lies in its capability of introducing additional spatial signal processing dimensions [1]. To date, a number of efficient signal processing techniques have been proposed to combat the harmful effect of wireless channel fading by adapting the channel passively. However, without the capability of modifying the electromagnetic environment actively, the performance of these methods is still limited in some extreme channel scenarios. To address this issue, researchers have proposed the concept of reconfigurable MIMO, aiming to actively optimize channel quality by altering the electrical characteristics of the antenna array. [2], [3], [4] discussed realizing reconfigurable MIMO by altering the antenna element spacing. However, the inter-antenna coupling of reconfigurable MIMO still limits its performance. In this sense, the emerging reconfigurable antennas (RA) make it possible to break through this bottleneck.