Conventional tracking filters incur mean tracking errors in the presence of a pilot-induced target maneuver. Chan has proposed a solution to this problem which uses the mean deviation of the residual innovation sequence to make correction to the Kalman filter. Bogler has improved the Chan's algorithm for the case of a one-dimensional Kalman filter by means of closed-form recursive relations and multiple hypothesis models. The technique of Bogler, while superior to the Chan's input estimation algorithm, requires a large number of filters to run in parallel(more than 20 filters are used in the simulation presented in [1]), so the computational load is heavy.
The purpose of this paper is to incorporate Bogler's algorithm into a realistic three-dimensional, coupled measurement, target tracker using decoupled target model, and to reduce computational load by modification of the Bogler's algorithm. Simulation results show that the proposed algorithm performs similar to the Bogler's algorithm, while its computational load is less than the Bogler's algorithm.