Track Related To Measurement Noise Filtering

A white noise observation model is widely used in tracking problem formulation. In practice, the measurement noise may not be white. And in some conditions, the observation model noise is correlated in direction due to indirect observation model. This correlated noise may degrade the performance of a tracking algorithm. Besides, the model error and computational error exit in practice, so Kalman filter operation is sometimes rendered totally unacceptable.In this paper, on actual project background, we study target tracking with correlated noise. New filters are given. Good tracking effect is demonstrated via digital simulation. The following is the main content of the paper:l.In the beginning of the paper, the target being researched about is declared to be the planes that are in cruising state, then the movement model and observation model of the target are proposed.2. Kalman filter and limited memory filter are described, and a new filter called inertial filter is given, to estimate the velocity of the target.3. The statistic traits of the measurement noise as an AR process is studied. With the state augmentation approach and the measurement difference approach, the effect of colored noise can be alleviated.4. Finally, the tracking performances in four noise conditions are compared via practical data, and parameter condition is analyzed in which the colored noise can be seen as white. Besides, inertial filter improves the estimate performance of the target velocity.