FDA‐MIMO Range and Angle Estimation via Unitary CP Decomposition With Plug‐and‐Play Denoising

This letter proposes a joint data‐model‐driven framework for range and angle estimation. By integrating a unitary CP decomposition with a plug‐and‐play deep denoiser, we develop an iterative scheme that estimates factor matrices under the physical structural constraints and denoising priors. This approach leverages the interpretability of tensor decomposition alongside the powerful denoising capability of deep networks in the real‐valued domain. Simulations show that the proposed method significantly enhances the parameter estimation accuracy compared to traditional CP decomposition and subspace methods.