Enhancing Channel Estimation in mmWave RIS-assisted SIMO-OFDM Systems: A Data-aided Expectation Maximization Approach

Channel Estimation (CE) in Reconfigurable Intelligent Surface (RIS)-assisted communication systems is crucial for reflection design at the RIS node and detection at the destination. However, CE for RIS-assisted systems generally requires numerous pilots to provide sufficient Degrees of Freedom (DoF) due to the lack of A/D and D/A converters at the RIS node and the high-dimensional space-time-frequency received signals. In this paper, we study a channel estimator with insufficient pilots for RIS-assisted mmWave Single-Input Multiple-Output Orthogonal Frequency Division Multiplexing (SIMO-OFDM) systems. Specifically, we explore the DoFs resulting from the data to conduct CE using the Maximum Likelihood (ML) estimator with the Expectation Maximization (EM) algorithm. Simulation results demonstrate that the proposed data-aided EM approach with fewer pilots significantly improves CE performance for the RISassisted SIMO-OFDM system without the need for additional pilots at the transmitter.