Target Localization for IRS-Assisted Massive MIMO Systems

Intelligent reflecting surface (IRS) becomes a new promising candidate technology for the next generation wireless communications. When the radar sensing is integrated into the IRS-assisted massive multiple-input multiple-output (MIMO) systems, it is challenging to detect targets-of-interest for the radar sensing, because the target echo signals could be also reflected by IRS, resulting in a fake angle issue for the target localization. To handle the fake angle issue, in this article, we propose a new sparse Bayesian learning (SBL)-based approach for the target localization with the IRS-assisted massive MIMO system. Taking into account the fact that the positions of IRS and base station (BS) remain unchanged, we novelly utilize the angular reciprocity between the downlink and uplink channels to facilitate the target localization. Then, we present a new hybrid sparse Bayesian framework to characterize the sophisticated structured sparsity brought by the angular reciprocity, where different hierarchical distributions are assigned to the target echo signals and uplink communication signals so as to enforce much sparser common structure corresponding to the fake angles. Finally, we exploit an expectation-maximization (EM)-based algorithm to jointly recover the common and independent sparse signals, which will identify the fake and true angles automatically. Simulation results verify the superiority of the proposed method.