Leveraging location and RIS elements grouping for the broadband channel estimation of RIS-aided mmwave massive MIMO

Abstract

Reconfigurable intelligent surface (RIS) is regarded as a promising technology to control intelligently the wireless environment. However, accurate and low-overhead channel estimation (CE) is considered a crucial challenge due to the large dimension of passive elements for RIS. This paper investigates the broadband CE of RIS-aided millimeter-wave (mmWave) communication system and proposes an atomic norm minimization (ANM)-based CE solution. Specifically, we propose a signal transmission protocol based on between base station (BS)-user equipment (UE) direct channel and BS-RIS-UE cascaded channel and design the pilot signals based on the prior location information. We utilize RIS elements grouping method to decrease pilot overhead and employ RIS beam broadening to mitigate the impact of elements grouping on the degrees of beamforming freedom. In direct CE, the decoupled atomic norm minimization (DANM) algorithm is exploited to estimate the channel angle parameters as well as the alternating direction method of multipliers (ADMM) algorithm is utilized to decline the algorithm complexity to satisfy the requirement of real-time signal processing. In cascaded CE, in order to improve the accuracy of CE, we obtain the channel angle parameters by employing an ANM problem with multiple measurement vectors (MMV). To theoretically evaluate the performance of the proposed algorithm, we derived the performance limits of mean square error (MSE) by Cramér-Rao lower bound (CRLB) analyses. Simulations results show that the proposed algorithm can succeed efficient CE under low-overhead circumstances.