A Proposed Channel Estimation Framework for Reconfigurable-Intelligent-Surface-Aided MIMO Communication System

Abstract

Reconfigurable intelligent surface (RIS) is considered a promising technology for dense Internet of Things (IoT) networks to enhance spectral efficiency and coverage due to its lower energy consumption and decreased hardware cost. However, channel estimation is a significant challenge for two reasons; first, RIS reflects signals without any signal processing capabilities, and second, it contains large numbers of reflected elements, making the pilot overhead extremely high. In this paper, we propose a two-phased RIS channel estimation framework for MIMO uplink transmissions with lower pilot overhead. In phase I, we take advantage of the fact that RIS elements reflect all IoT devices through the same BS-RIS channel and propose a dual-link time division duplex (TDD) pilot transmission scheme for estimating the BS-RIS channel where the pilots are transmitted from BS to RIS and then reflected back to BS. Afterward, we propose an algorithm to obtain the BS-RIS channel from the dual link BS-RIS-BS cascaded channel. In phase II, we estimate the user channels between RIS and devices based on the knowledge of the BS-RIS channel obtained in phase I. Simulation results indicate that the proposed framework attained a lower normalized mean square error (NMSE) and higher sum rate than other channel estimation frameworks with lower pilot overhead and complexity overhead.