Explainable AI-Based Channel Estimation For Intelligent Reflecting Surface-Aided Networks

Intelligent Reflecting Surface (IRS)–aided wireless networks have emerged as a key enabling technology for next-generation 6G communications due to their ability to intelligently reconfigure the wireless propagation environment. Accurate channel state information (CSI) is critical for realizing the full potential of IRS-assisted systems; however, channel estimation becomes highly challenging because of the large number of passive reflecting elements, cascaded channel structure, and dynamic wireless conditions. To address these challenges, this paper proposes an Explainable Artificial Intelligence (XAI)–based channel estimation framework for IRS-aided wireless networks. The proposed approach leverages a deep learning model to capture the nonlinear relationship between pilot signals and cascaded IRS channels, while integrating explainability mechanisms to provide transparent and interpretable insights into the estimation process. XAI techniques such as feature attribution and attention-based analysis are employed to identify the most influential IRS elements, pilot configurations, and environmental parameters contributing to accurate CSI prediction. Simulation results demonstrate that the proposed XAI-based model significantly outperforms conventional estimation methods in terms of normalized mean square error (NMSE) and robustness under varying signal-to-noise ratios, while offering enhanced interpretability and reliability. The proposed framework bridges the gap between high-performance AI-driven channel estimation and trustworthy wireless communication systems, making it a promising solution for future intelligent and adaptive 6G networks.