Location-Aware Communication for RIS-Aided Distributed MIMO Systems

Abstract

In reconfigurable intelligent surface (RIS)-aided distributed multiple-input multiple-output (D-MIMO) systems, instantaneous channel state information (CSI) is essential for the optimization of the RIS phase shifts. However, the D-MIMO systems often involve numerous RISs and access points (APs), requiring the transmission of a significant amount of pilots for CSI estimation, thus resulting in high signaling overhead. To mitigate the high signaling overhead associated with CSI estimation, this paper introduces a location-aware communication method that only relies on large-scale fading information. Two new challenges have arisen in the proposed location-aware communication method. First, it is challenging to design a user position estimation method to achieve fast and accurate localization. Second, it is hard to handle the location uncertainty brought by the estimation error when performing beamforming. In response to the above challenges, we first design a space-time cooperative beam training-based localization method to achieve a favorable trade-off between localization accuracy and time cost. Subsequently, to handle the location uncertainty, by taking historical error distribution into consideration, we formulate a distributionally robust optimization (DRO)-based problem for sum-rate maximization. Numerical evaluations demonstrate the effectiveness of our proposed localization method, which can achieve sub-meter accuracy with low beam training overhead. Furthermore, the proposed DRO-based beamforming method can improve the sum-rate by 11.4% compared to the robust optimization and non-robust schemes.