Title:Low-complexity Joint Beamforming for RIS-Aided Multi-User Downlink over Correlated Channels

Abstract:This paper considers the reconfigurable intelligent surface (RIS)-assisted multi-user communications, where an RIS is used to assist the base station (BS) for serving multiple users. The RIS consisting of passive reflecting elements can manipulate the reflected direction of the incoming electromagnetic waves by adjusting the phase shifts of the reflecting elements. Alternating optimization (AO) based approach is commonly used to determine the phase shifts of the RIS elements. While AO-based approaches have shown significant gain of RIS, the complexity is quite high due to the coupled structure of the cascade channel from the BS through RIS to the user. In addition, the sub-wavelength structure of the RIS introduces spatial correlation that may cause strong interference to users. To handle severe multi-user interference over correlated channels, we consider adaptive user grouping previously proposed for massive mutli-input and multi-output (MIMO) systems and propose two low-complexity beamforming design methods, depending on whether the grouping result is taken into account. Simulation results demonstrate the superior sum rate achieved by the proposed methods than that without user grouping. Besides, the proposed methods can perform similarly to the AO-based approach but with much lower complexity.