Blind channel estimation for wideband RIS-assisted mmWave multi-user system with direct channels using structured subspace

In this paper, we propose a novel blind channel estimation framework for reconfigurable intelligent surface (RIS)-assisted millimeter wave (mmWave) wideband multi-user multiple-input single-output (MU-MISO) systems. Specifically, a joint cascaded and direct channel estimation approach is developed based on a structured subspace method. The proposed technique exploits the inherent Toeplitz structure of the channel matrix to formulate a cost function, which is subsequently minimized to recover accurate channel estimates. Unlike conventional methods, the proposed approach operates in a completely blind manner without relying on pilot signals, thereby conserving bandwidth while enhancing both spectral efficiency and overall system throughput. Simulation results are provided to illustrate the attractive benefits of the proposed method.