Localization and Channel Reconstruction for Extra Large RIS-Assisted Massive MIMO Systems

Reconfigurable intelligent surface (RIS) is a promising material that can passively manipulate electromagnetic waves and improve the quality of mobile communication services at a low cost. It can be made large to extend the service region and acquire the ability for localization enhancement. However, the lack of a signal processing module at the RIS makes channel estimation a difficult problem. When employing an extra large RIS, users are in the near field of the RIS and the channel shows spatial nonstationarity, making the problem more complicated. In this paper, these challenges are addressed by a low-overhead joint localization and channel reconstruction scheme proposed for extra large RIS-assisted massive multi-input multi-output systems. This scheme can accurately identify the visibility region (VR) of each user, find the user positions by exploiting the near-field characteristics, and reconstruct the channels by jointly utilizing the pilots of multiple users. Numerical results demonstrate that the identified VR covers more than 97% of the real VR, and the user localization accuracy reaches centimeter level at the millimeter-wave frequency band. A more accurate channel reconstruction result than that of existing works can also be obtained. This work verifies the great potential of RIS and is a solid step toward the integration of communication and sensing.

I. Introduction

Reconfigurable intelligent surface (RIS) is a passive material that can flexibly control electromagnetic waves and improve electromagnetic propagation environment [1]. Owing to its low-cost advantage, RISs can be densely deployed to recover services at mobile communication holes and enhance signal strength, which is a more economic option compared with some active nodes such as relays. By providing controllable supplementary links, RIS can improve localization accuracy, thereby achieving a good integration of communication and sensing. Given these attractive advantages, RIS is regarded as a potential key technology for sixth-generation mobile communications [2], [3]. Extensive works have studied RIS-assisted mobile communications, including theoretical analysis [4]–[8] and experimental measurements [9]–[11], and the results provide useful guidance for RIS-assisted transceiver design. Notably, the optimal design always relies on the acquisition of channel state information (CSI). CSI also supports the localization of the user. Therefore, to harvest gains in communication and sensing, channel estimation in RIS-assisted mobile communication systems is essential.