Joint DFO and channel estimation scheme for RIS assisted OFDM systems

Abstract

In the reconfigurable intelligent surface (RIS) assisted mobile wireless communications, the performance gains provided by RIS cannot be achieved without the accurate channel estimation. Many time-varying channel estimation algorithms have been proposed for such systems, but none of them consider the effect of Doppler frequency offset (DFO) synchronization error, which can seriously degrade the accuracy of channel estimation, especially in the Rician fading channel for the orthogonal frequency division multiplexing (OFDM) systems. In our work, the effect of DFO synchronization error on the time-varying channel estimation is firstly analyzed, and then a deep learning based joint DFO and channel estimation scheme is proposed. Specifically, the threshold-based noise reduction algorithm and the dual convolution neural networks (CNN) are applied for improve the estimation accuracy. To enhance the practicability of the proposed method, the training target of CNN network is set as the actual estimation value with high accuracy, rather than the ideal information. The simulation results show that the proposed method exhibits superior performance compared to the existing algorithms, and it has an acceptable computational complexity.