RCS–Doppler-Assisted MM-GM-PHD Filter for Passive Radar in Non-Uniform Clutter

In passive radar, the multiple model probability hypothesis density (MM-PHD) filter has demonstrated robust capability in tracking multi-maneuvering targets. Nevertheless, non-uniform clutter in practical scenarios causes misestimation of component weights, thereby generating false targets. To solve the false targets problem, a feature-matching MM-PHD (FM-MM-GM-PHD) algorithm for passive radar tracking is proposed in this paper. First, the measurement likelihood function was refined by leveraging target radar cross-section (RCS) and Doppler features to assist in suppressing false targets and reduce clutter interference. Additionally, the proposed algorithm incorporated adaptive component pruning and absorption processes to enhance tracking accuracy. Finally, a missed-alarm correction mechanism was introduced to compensate for measurement losses. Simulations of the passive radar results validated the findings that the proposed algorithm outperformed the traditional MM-PHD filter in both tracking accuracy and cardinality estimation. This superiority was particularly pronounced in non-uniform clutter environments under low detection probabilities.