Masking Effect Mitigation for FM-Based Passive Radar via Nonlinear Sparse Recovery

Nowadays, most passive radars use cross-ambiguity function (CAF) computation for signal processing. Although it is easy to implement, the performance may degrade significantly due to the masking effect among multiple echoes when the waveform's ambiguity function (AF) is not good. To mitigate the masking effect and relieve the waveform requirement, we propose the idea of nonlinear signal processing. Specifically, we formulate the passive radar signal processing into an estimation problem. Nonlinear signal processing is introduced by exerting nonlinear regularization to the estimation model, which makes it resemble a sparse problem. First, we analyze the strong masking effect in frequency modulation (FM) broadcast signals. And then, an efficient alternating direction method of multipliers (ADMM) based algorithm is proposed to solve the large-scale sparse recovery problem in passive radar. Furthermore, we propose a criterion for setting the regularization parameter and derive upper and lower bounds of the regularization parameter based on its mathematical and physical meaning. We also provide a measure to mitigate model mismatch due to delay-Doppler gridding. Finally, simulations and field experiment results demonstrate the practical feasibility of the proposed algorithm for masking effect mitigation in passive radars.