Sparsity Channel Estimation for Reconfigurable Intelligent Surface Aided MIMO Systems

This paper investigates the performance of cascaded channel estimation in the millimeter-wave multiple-input multiple-output (MIMO) systems via a reconfigurable intelligent surface (RIS). The main goal is to design a low-overhead cascaded channel estimation scheme that provides for a good performance of wireless networks deployed in hot spots. However, the high-dimensional channel of RIS links and the passive feature of RIS without signal processing capability make the acquisition of channel state information a challenging, and thus, channel estimation in RIS-assisted wireless communication systems requires high pilot overhead. In the practical scenario, there are limited scatters around the base station and the RIS. The angular cascaded channel has a few non-zero elements, which exhibit the sparsity. Benefiting from these special channel characteristics, we propose a Sparsity Generalized Orthogonal Matching Pursuit based cascaded channel estimation scheme by integrating the structured sparsity into the GOMP algorithm to reduce pilot overhead. the proposed solution improves estimation performance by approximately 2 dB, while reducing average running time to a level comparable to the Oracle LS scheme. Meanwhile, our scheme incurs a pilot cost decrease of approximately 14:3% compared to traditional approaches, when the NMSE is -2 dB. The proposed scheme sheds a light to engineers working in wireless communication on enhancing the channel estimation performance of existing wireless systems.