Parallel Filtering Algorithm in Three Dimensions for IMM Maneuvering Target Tracking

Abstract

According to the physical truths those are the complexity of special target maneuvering and the inconformity of maneuvering degrees in three dimensions, the problems of model mismatching and inaccuracy in traditional IMM were analysed, then a parallel filtering algorithm in three dimensions for IMM maneuvering target tracking is presented. The model set of this algorithm consists of the CV and the modified CS model, which can adaptively tracking target under different maneuvering levels; the parallel IMMs in three dimensions can update model probabilities respectively according its maneuvering reality,which ensures the accuracy of model probabilities. The simulation results indicate that the proposed algorithm gets higher tracking precision and decrease 1/3 computational complexity than traditional IMM. That is to say, it has a good practical prospect in maneuvering target tracking in space.