A Two-Stage Adaptive Algorithm for Tracking Multi-mode Maneuvering Targets

This paper proposes an adaptive algorithm for tracking multi-mode maneuvering targets using bearing and frequency measurements in two stages: tracking establishment and tracking maintenance. The algorithm uses a three-model structure with ceiling, flooring and adaptive models to handle different levels of target maneuvers. The posterior probability of each model is calculated after one update and the adaptive model can adjust system noise coefficient according to it. The final estimation is a data fusion of the results from the three models. By testing on a simulated scenario where the target performs three consecutive turns with increasing curvature, it is proved that the algorithm can effectively track the target under various situations.