USING TWO-STAGE KALMAN FILTERS AS OBSERVERS FOR SIMULTANEOUS TRAJECTORY TRACKING AND UNKNOWN INPUT ESTIMATION

Target tracking with an unknown bias is important in the fieldsof navigation, trajectory determination, and so on. Kalman filter issimple in principle and widely used, however, it has poor ability forstate estimation with unknown bias. To improve its performances,robust two-stage Kalman filter (RTSKF) and optimal two-stageKalman filter (OTSKF) are adopted for target tracking and unknownbias estimation and their behaviours are compared comprehensivelyin this work. First, the feasibility of these filters is validated undera low-noise environment. Then, the robustness to resist the noises isinvestigated when the intensities of process noise and measurementnoise are changed from 0.01 to 0.50. Results demonstrate theOTSKF has stronger robustness than RTSKF. Finally, the types ofunknown bias are changed to test the flexibility and accuracy totrack manoeuvring target.