Extended Target Localization in Distributed MIMO Radars With Low-Resolution ADCs: A Diffusion-Aware Bayesian Approach

Distributed MIMO radars with widely separated antennas offer spatial diversity and enhanced parameter estimation capabilities. The use of low-resolution ADCs significantly reduces data volume and system costs. However, most existing research focuses on point-target models, neglecting the complexities of extended targets, which consist of multiple closely spaced scatterers. In this paper, we address the localization of extended targets in low-bit distributed MIMO radars. The problem is formulated as a multi-task sparse recovery model, characterized by a two-level block sparsity pattern. To leverage this structure, we extend the pattern-coupled sparse Bayesian learning (PCSBL) algorithm to low-bit quantization scenarios. Furthermore, to address the limitations of the diffusion-ignorant sub-optimal solution used in the original PC-SBL algorithm, we propose a diffusion-aware hyperparameter estimation approach to derive an analytical optimal solution, ensuring enhanced reconstruction accuracy and stable convergence. Numerical experiments validate that our proposed methods achieve superior localization accuracy compared to existing state-of-the-art low-bit sparse recovery algorithms.