A hybrid Fuzzy dynamic model for maneuvering targets

Accuracy and precision of estimated state of target depends on the used motion model in tracking algorithm. Motion model of maneuvering targets is changed uncertainly, in each instant for their varying acceleration. In this paper, to minimize the estimation error produced from using mismatch motion model, an adaptive motion model is proposed. It is obtained based on adjusting the maneuvering time constant of the Singer model according to the output of a fuzzy ARTMAP which classifies motion into constant velocity and constant acceleration in each step time. The improved tracking performance of using the proposed motion model over using popular models such as the interacting multiple model and the Singer model, is illustrated by simulation in tracking of an anti-ship missile.