Deep Learning for Super-Resolution Channel Estimation in Reconfigurable Intelligent Surface Aided Systems

Reconfigurable intelligent surface (RIS) enables the configuration of propagation environment. Channel estimation is an essential task in realizing RIS-aided communication system. A multi-user multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) system involves cascaded channels with high dimensions and sophisticated statistics. Thus, implementing optimal minimum mean square error (MMSE) integration computation infeasible practice. To accurately estimate accuracy a MIMO-OFDM system, we model channel state information (CSI) as image super-resolution (SR) problem to recover denoise matrix. Particularly, convolutional neural network based on (SRCNN) denoising (DnCNN), named SRDnNet, then proposed. By taking estimated at pilot positions low-resolution image, enhanced SRCNN can fully exploit features inputs learn suitable interpolation method generate coarse The DnCNN element-wise subtraction structure additive noise coefficients from simulation results demonstrate effectiveness excellent performance proposed SRDnNet.