Multiple Residual Dense Networks for Reconfigurable Intelligent Surfaces Cascaded Channel Estimation

TL;DR

This paper introduces three residual neural network models leveraging low rank structures for accurate channel estimation in reconfigurable intelligent surface systems, demonstrating superior performance over benchmarks.

Contribution

It proposes three novel residual neural network architectures specifically designed for RIS channel estimation, exploiting low rank properties for improved accuracy.

Findings

All three models outperform existing benchmark schemes.

The residual dense network shows the best estimation accuracy.

Simulation results indicate effective evolution and potential of the proposed methods.

Abstract

Reconfigurable intelligent surface (RIS) constitutes an essential and promising paradigm that relies programmable wireless environment and provides capability for space-intensive communications, due to the use of low-cost massive reflecting elements over the entire surfaces of man-made structures. However, accurate channel estimation is a fundamental technical prerequisite to achieve the huge performance gains from RIS. By leveraging the low rank structure of RIS channels, three practical residual neural networks, named convolutional blind denoising network, convolutional denoising generative adversarial networks and multiple residual dense network, are proposed to obtain accurate channel state information, which can reflect the impact of different methods on the estimation performance. Simulation results reveal the evolution direction of these three methods and reveal their superior performance compared with existing benchmark schemes.