Abstract:In the realm of reconfigurable intelligent surface (RIS)-assisted communication systems, the connection between a base station (BS) and user equipment (UE) is formed by a cascaded channel, merging the BS-RIS and RIS-UE channels. Due to the fixed positioning of the BS and RIS and the mobility of UE, these two channels generally exhibit different time-varying characteristics, which are challenging to identify and exploit for feedback overhead reduction, given the separate channel estimation difficulty. To address this challenge, this letter introduces an innovative deep learning-based framework tailored for cascaded channel feedback, ingeniously capturing the intrinsic time variation in the cascaded channel. When an entire cascaded channel has been sent to the BS, this framework advocates the feedback of an efficient representation of this variation within a subsequent period through an extraction-compression scheme. This scheme involves RIS unit-grained channel variation extraction, followed by autoencoder-based deep compression to enhance compactness. Numerical simulations confirm that this feedback framework significantly reduces both the feedback and computational burdens.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that in the reconfigurable intelligent surface (RIS) - assisted communication system, due to the fact that the cascaded channels between the base station (BS) and the user equipment (UE) have different time - varying characteristics, resulting in excessive channel feedback overhead. Specifically: 1. **Challenges in channel estimation**: Since the positions of the BS and RIS are fixed while the UE is mobile, this makes the BS - RIS channel and the RIS - UE channel exhibit different time - varying characteristics. This difference makes it difficult to estimate these two channels separately and increases the feedback overhead. 2. **Reduction of feedback overhead**: In order to reduce the feedback overhead, an effective framework is required to capture the time - varying characteristics of the cascaded channels. Traditional channel feedback methods have very large feedback overhead when dealing with high - dimensional channels in large - scale MIMO systems, especially in the case of lack of channel reciprocity in frequency - division duplex (FDD) systems. For this reason, the paper proposes a two - time - scale cascaded channel feedback framework based on deep learning, aiming to solve the problem in the following ways: - **Extraction and compression scheme**: This framework introduces an extraction - compression scheme for generating a compact representation of cascaded channel changes. Specifically, it includes change extraction at the RIS unit level and deep compression based on auto - encoders. - **Two - time - scale feedback**: Feedback the entire cascaded channel on a larger time scale and feedback the changes in the cascaded channel (i.e., the ratio vector) on a smaller time scale. This can effectively reduce the feedback overhead and computational complexity. Through this method, the new framework proposed in the paper can significantly reduce the feedback overhead and computational burden while maintaining good channel estimation performance. ### Summary of mathematical formulas - **Representation of cascaded channels**: \[ H(t)=\begin{bmatrix} a_1(t)b_1(t)\\ \vdots\\ a_{N_{\text{RIS}}}(t)b_{N_{\text{RIS}}}(t) \end{bmatrix} \] where \(a_i(t)\) is the \(i\)-th element of the RIS - UE channel, and \(b_i(t)\) is the \(i\)-th row of the BS - RIS channel. - **Calculation of ratio vectors**: \[ p_i(t)=\frac{h_i^H(0)h_i(t)}{\|h_i(0)\|_2^2} \] where \(p_i(t)\) is the \(i\)-th element of the ratio vector, representing the cascaded channel change corresponding to the \(i\)-th RIS unit. - **Calculation of feedback overhead**: \[ N_{\text{equ}}=\frac{N_QN_{s1}+TN_{s2}}{T + 1} \] where \(N_Q\) is the number of quantization bits, \(N_{s1}\) and \(N_{s2}\) are the codeword dimensions of AE1 and AE2 respectively, and \(T\) is the number of time intervals on the small time scale. These formulas and methods together form the core contributions of the paper, solving the problem of excessive channel feedback overhead in RIS - assisted communication systems.

Efficient Deep Learning-based Cascaded Channel Feedback in RIS-Assisted Communications

Deep Learning-Based Joint Channel Estimation and CSI Feedback for RIS-Assisted Communications

Deep Learning-based CSI Feedback for RIS-assisted Multi-user Systems

Time-Varying Channel Prediction for RIS-Assisted MU-MISO Networks Via Deep Learning

Channel Feedback for Reconfigurable Intelligent Surface Assisted Wireless Communications.

RISNet: A DL-Based High Accuracy CSI Feedback Approach for RIS-Aided FDD Systems

Deep Learning-Based Channel Estimation for Double-RIS Aided Massive MIMO System

Deep Learning-Based Time-Varying Channel Estimation for RIS Assisted Communication

Federated Learning Enabled Channel Estimation for RIS-Aided Multi-User Wireless Systems

Parallel Factor Decomposition Channel Estimation in RIS-Assisted Multi-User MISO Communication

Deep Learning for Secure Transmission in Reconfigurable Intelligent Surface-Assisted Communications.

Joint Channel Estimation and Signal Recovery in RIS-Assisted Multi-User MISO Communications.

A Robust Deep Learning-Based Beamforming Design for RIS-Assisted Multiuser MISO Communications with Practical Constraints

Joint Channel Estimation and Signal Recovery for RIS-Empowered Multi-User Communications

RISMCNet: A Two-Timescale CSI Feedback Solution for RIS-aided Multi-Carrier systems

Reconfigurable Intelligent Surface-Assisted Aerial-Terrestrial Communications Via Multi-Task Learning

Two-Stage Channel Estimation for Hybrid RIS Assisted MIMO Systems

Secure Energy Efficiency in Communication Systems Based on Deep Learning

A Deep Learning-Based Low-Overhead Beam Tracking Scheme for Reconfigurable Intelligent Surface-Aided Multiple-Input and Single-Output Systems with Estimated Channels

Channel Customization for Limited Feedback in RIS-assisted FDD Systems

Mmwave RIS-Assisted SIMO Channel Estimation Based on Global Attention Residual Network