Uneven Sensor Bias Removal for the Track-to-Track Association Problem

In real distributed multi-radar multi-target tracking systems, uneven sensor bias might bring great challenge to track association because it would make an uneven rotation of all the radar tracks, which always causes association mistakes. Unfortunately, correcting all the tracks in the traditional method of using a single average estimation of sensor bias would lead to false tracks in some regions. Unlike most of the published previous works, this paper for the first time proposes an adaptive technique for track-to-track association against uneven sensor bias. The algorithm consists of three stages: the adaptive track bias analysis, the online estimation of uneven sensor bias and the adaptive association adjustment with bias removal. We also present an anti-bias track association flow. The simulation results show the effectiveness of the technique, which could accurately remove the uneven sensor bias and significantly improve the association probability, suggesting a great value in engineering.