Target Motion Analysis Using Single Sensor Bearings-Only Measurements

The stipulation that measurements are obtained from only a single sensor adds additional complexity to the bearings-only tracking (BOT) problem since target states can then be observed only after a sensor platform maneuver. Because of lack of target knowledge from measurements and highly nonlinear between target state and measurements, filter initialisation is difficult and the consequent tracker instability is an inherent issue. In this paper, we discuss the use of progressive correction technique to improve the extended Kalman filter (EKF) performance for BOT problem and recall a pseudo linear least square (PLLS) filter which is derived from a batch of measurements and does not require initialisation. The performance of both filters are analysed in a benchmark scenario which was used for demonstrating the performance of a range parameterised EKF implemented in the modified polar coordinates (MPC). Simulation results demonstrated the robustness and effectiveness of these simple techniques. Index Terms—Bearings-only tracking, Pseudo linear least square estimator, Filter initialisation, Nonlinear filtering, Pro- gressive correction.