Pyung-Soo Kim
Target tracking, IMM, FIR filtering, Kalman filtering
This paper proposes a computationally efficient approach for the target tracking of two-dimensional target systems with unknown maneuvers. To describe efficiently target with maneuvers, only two target models are developed. To estimate target trajectories and unknown maneuvers, two tracking filters using only the most recent finite observations are used due to fast estimation performance and low computation burden. It will be shown that constant-bias maneuvers do not induce the error in estimates of target trajectories. A detection method is also developed in order to indicate the presence of maneuvers and select the valid estimate from two filters. It is shown that the on-line computation amount of the proposed approach is much less than that of the well-known interacting multiple model (IMM) approach. Via extensive simulations, the performance of the proposed target tracking approach is evaluated by the comparison with existing IMM approaches.
Important Links:
Go Back
