In recent years there has been a growing interest in reconfigurable intelligent surfaces (RISs) as enablers for the realization of smart radio propagation environments which can provide performance improvements with low energy consumption in future wireless networks. However, to reap the potential gains of RIS it is crucial to jointly design both the transmit precoder and the phases of the RIS elements. Within this context, in this paper we study the use of multiple RIS panels in a parallel or multi-hop configuration with the aim of assisting a multi-stream multiple-input multiple-output (MIMO) communication. To solve the nonconvex joint optimization problem of the precoder and RIS elements targeted at maximizing the achievable rate, we propose an iterative algorithm based on the monotone accelerated proximal gradient (mAPG) method which includes an extrapolation step for improving the convergence speed and monitoring variables for ensuring sufficient descent of the algorithm. Based on the sufficient descent property we then present a detailed convergence analysis of the algorithm which includes expressions for the step size. Simulation results in different scenarios show that, besides being effective, the proposed approach can often achieve higher rates than other benchmarked schemes.
翻译:近年来,人们日益关注可重新配置的智能表面作为实现智能无线电传播环境的促进因素,这种环境可以提高未来无线网络低能消耗的性能;然而,为了获得RIS的潜在收益,必须联合设计传输预编码器和RIS元素的各个阶段;在此背景下,我们研究如何以平行或多窗口配置的方式使用多个RIS面板,以协助多流多投入多产出通信;为了解决前编码器和RIS元素的非康韦克斯联合优化问题,目的是最大限度地实现可实现的速率,我们提议采用基于单调加速准偏差法的迭代算法,其中包括一个外推法步骤,以提高趋同速度和监测变量,确保算法充分脱落。基于充分的血统属性属性,我们随后对算法进行详细的趋同分析,其中包括步骤大小的表达。模拟结果显示,除了有效外,拟议的方法往往可以达到高于其他基准计划的比率。