Active Element Selection for Channel Estimation in RIS-Aided Wireless Communication Systems

The hybrid reconfigurable intelligent surface (RIS) architecture consisting of both active elements (AEs) and passive elements (PEs) is promising for use in milimeter wave wireless communication systems. In this paper, we propose a novel scheme for selection of AEs in this kind of RIS-aided system to enhance the channel estimation performance. The RIS is first partitioned into multiple subregions and an appropriate number of AEs are then selected strategically from each subregion according to some criteria. A greedy coordinate descent algorithm is also proposed to reduce the computational complexity of the selection process. With the selected AEs, channel estimation is performed between the user equipment and RIS based on compressed sensing. Computer simulation results demonstrate that the proposed AE selection scheme induces a significant performance improvement in channel estimation over the random AE selection approach.