Channel Estimation in RIS-aided Multi-User OFDM Communication Systems

Reconfigurable Intelligent Surfaces (RIS) are planar arrays constructed of independently controllable passive reflecting elements that change the phase and the amplitude of the reflected signals between Base Station (BS) and User Equipment (UE). To facilitate the coherent combination of the reflected signals, the Channel State Information (CSI) of the channels should be known by the communication network. The main issue with CSI estimation in unstructured channels is the high number of channels and the long training time that is required to apply orthogonal training reflection patterns at the RIS. With multiple users, the required training period becomes even longer. Therefore, this study proposes allocating the users with orthogonal pilot sequences, which are transmitted over the orthogonal frequency division multiplexing (OFDM) frequency domain with the aim of separating the user signals in the frequency domain. In the study, the cascaded channels are separated in the time domain by a reflection pattern consisting of a DFT matrix and the users are separated in the frequency domain by the use of the Hadamard matrix. The algorithms are simulated over a 3GPP-defined Tapped Delay Line (TDL) channel model as well as a Rayleigh fading channel. The results show that due to the use of orthogonal sequences, channel estimation improves in flat fading channels, while the performance is also of acceptable level in frequency selective channels.