Filter for tracking a maneuvering target in a mixed coordinate system

A Kalman filter algorithm in the spherical-rectangular coordinate systems for tracking a maneuvering target was introduced. Compared with the conventional approach, the main improvements are: (1) an extended state vector is proposed to improve the state estimation (i.e. position, velocity and acceleration) even when the sensor's data are color contaminated; (2) a mix rectangular-spherical coordinate system is selected to model the target motion and position measurements from sensor thus avoiding the indescribable problem of measurement noise; (3) the Kalman filter equations are fully decoupled through using the rotational coordinate system and the filter gain rotation algorithm. The calculated results show that there is significant improvement in tracking capability over the methods discussed by other researchers.