Abstract
Research summary
Identifying targets as belonging to unmanned aerial vehicles (UAVs) or birds exploiting radar track sequences is critical for flight operations management and airport critical infrastructure safety. In this context, a hierarchical classification framework with multi-feature trustworthy integration (MFTI), multi-preference decision fusion (MPDF), and multi-window negotiation-based iterative reasoning (MWNIR) is proposed. The MFTI employs a short-term multi-branch deep evidential neural network, which leverages the Dirichlet process to reliably integrate the motion and radar cross-sections (RCS) features of the target. The MPDF combines the outputs of two MFTIs with different preferences based on the Dempster-Shafer theory, providing reliable classification results on short-term track segments with different qualities. Finally, to reduce the impact of track data imprecision on the inference process, the MWNIR exploits a negotiation and iterative temporal Dempster-Shafer evidence fusion. Experiments on real-world radar track data demonstrate that the proposed algorithm can outperform some state-of-the-art methods in terms of classification accuracy and reliability.