A comparison of Markov and constant turn rate models in an adaptive Kalman filter tracker

The authors compare the performance of two different models of target acceleration when used with a tracking algorithm that includes measurement of target aspect angle. Both first-order Markov (FOM) and constant turn rate (CTR) target acceleration models are integrated into a tracking algorithm which uses a target image sensor in determining the orientation of highly maneuvering targets. Maneuvering target models for fixed wing aircraft and helicopters are considered. Both models of target acceleration are incorporated into the algorithm to examine the accuracy to which they can track and predict the trajectory of a maneuvering target as a function of both time and spatial location. The computational load associated with each acceleration model when incorporated into the algorithm is also considered. It is shown that the state estimator which considers target aspect angle and utilizes a first-order Markov model is superior to the CTR approach. The FOM approach also has relatively low computational requirements