Swarm target identification and orientation modality estimation using radar polarization features

UAV swarm enrich their applicability in civil and military fields. This elevates the necessity of radar surveillance for noncooperative swarm targets. Birds and UAVs have highly similar radar features and their swarm modalities bring challenges for accurate radar identification. This paper builds a feature space using radar polarization responses for swarm target identification. The relevance between target orientation modality and polarization features enables modality estimation using aspect ratio and canting angles. A target orientation modality estimation method supported by support vector regression model is proposed. Polarization features and regressor are embedded into a Tracking-While-Recognizing framework to achieve dynamic target identification. Target track and feature datasets are built to validate the method comprehensively. Compared with other radar identification solutions, the method has better classification capability in distinguishing single and swarm modalities. Target orientation modalities are estimated with acceptable numerical deviations. This enables refined swarm target identification within complex battlefield environments. The efficiency performance could support its application significance.