Near-Field Sparse Bayesian Channel Estimation and Tracking for XL-IRS-Aided Wideband mmWave Systems

The rapid development of 6G systems demands advanced technologies to boost network capacity and spectral efficiency, particularly in the context of intelligent reflecting surfaces (IRS)-aided millimeter-wave (mmWave) communications. A key challenge here is obtaining accurate channel state information (CSI), especially with extremely large IRS (XL-IRS), due to near-field propagation, high-dimensional wideband cascaded channels, and the passive nature of the XL-IRS. In addition, most existing CSI acquisition methods fail to leverage the spatio-temporal sparsity inherent in the channel, resulting in suboptimal estimation performance. To address these challenges, we consider an XL-IRS-aided wideband multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) system and propose an efficient channel estimation and tracking (CET) algorithm. Specifically, a unified near-field cascaded channel representation model is presented first, and a hierarchical spatio-temporal sparse prior is then constructed to capture two-dimensional (2D) block sparsity in the polar domain, one-dimensional (1D) clustered sparsity in the angle-delay domain, and temporal correlations across different channel estimation frames. Based on these priors, a tensor-based sparse CET (TS-CET) algorithm is proposed that integrates tensor-based orthogonal matching pursuit (OMP) with particle-based variational Bayesian inference (VBI) and message passing. Simulation results demonstrate that the TS-CET framework significantly improves the estimation accuracy and reduces the pilot overhead as compared to existing benchmark methods.