A Target Tracking Algorithm with Derivation Measurement

In this paper a new observation model is presented to improve the state estimation and prediction in a target tracking problem. Comparing with conventional approaches, the following are distinguished points of the approach. First, the measurement equation is set up in the polar coordinate and even combines the derivation measurement with the usual position measurements ( i.e. there are 6 sensor data now: range, azimuth, elevation angle, range rate, azimuth rate, and elevation rate). Second,the observation noise of sensor data is considered as a colored one and be set up as the model of AR(1), and by means of a pseudo measurement model, the requirement of Kalman filter will be satisfied. As a result, the accuracy of both the observation and the prediction will be increased.