Novel debiased converted measurement kalman filter algorithm

A novel debiased converted measurement Kalman filter algorithm was proposed, in which specific expression of mean and covariance for converted measurement Kalman noise were inferred by inserting an virtual item firstly, and then the optimal estimation of mean and covariance for converted measurement noise were deduced in the mean-square sense based on all the past measurements. The new algorithm was applied in tracking during missile-target encounter which required high filter accuracy, and Monte Carlo simulation was performed. The simulation results demonstrate the Root Mean Square Error of position and velocity are smaller, the accuracy is higher, the true estimation errors are quite consistent with the computed covariance of the filter, and the proposed filter is more credible. Besides, it keeps good performance even if the measurement noise becomes larger.