Beamforming and Tracking Assessment with Passive Radar Experimental Data
Tri–Tan Cao, Tuyet Vu, Marion Byrne, Nabaraj Dahal, Paul E. Berry, Ayobami B. Iji, Daniel Gustainis, Martin Ummenhofer, Michael Köhler · 2020
Using experimental data captured with a passive radar linear array built by Defence Science and Technology (DST) Group, the performance of various beamforming algorithms (beamformers) for direction-of-arrival (DoA) estimation is comparatively assessed. Beamformers' DoA estimates are passed to a tracker which performs angle tracking, and the resulting track quality is examined based on the Optimal Sub-Pattern Assignment Metric for Multiple Tracks (OSPAMT). By matching the ground-truth tracks with the confirmed tracks, the OSPAMT takes into account both the locality error (how close the confirmed tracks are to the ground-truth tracks) and the cardinality error (the numbers of false tracks, track breaks and missed tracks). The best OSPAMT score is achieved by a newly-designed sparse solution beamformer derived from a compressive sensing principle and based on off-grid approaches.