Integrated Trajectory Optimization and Cubature Kalman Filter for UAV-Based Target Tracking with Unknown Initial Position

This paper investigates the target localization/tracking by angle-of-arrival (AOA) sensors carried on an unmanned aerial vehicle (UAV). In many practical applications, the UAV's own position is not available, since the global position system (GPS) can hardly work in the indoor environment or interference region. Therefore, considering the unknown initial position of the UAV, a modified cubature Kalman filter (CKF) is developed to estimate both the target states and UAV's initial position jointly by leveraging a benchmark anchor. To further improve the estimation efficiency, we propose an algorithm to optimize the UAV flying trajectory by minimizing the the trace of the estimation covariance matrix in the CKF. According to the simulation results, an observation is found that the UAV will keeping flying alternatively between the anchor and target to guarantee the estimation performance.