Tracking Maneuvering Targets
You He, Xiu Jianjuan, Xin Guan · 2016
This chapter focuses on tracking maneuvering targets, or rather, the problem of uncertainty in establishing target model parameters in the filtering process. The single-model adaptive tracking algorithms for maneuvering targets include the modified input estimation, Singer model, current statistical model, and jerk model algorithms, while the multiple-model adaptive tracking algorithms include the multiple model and IMM algorithms. The chapter presents the jerk model algorithm, which, similarly to the Singer algorithm, models the process noise as the color noise, and needs to find the derivative of the acceleration in real time, that is, to estimate the acceleration. It discusses the interactive multiple model algorithm, a target tracking approach frequently used in recent years. The modified current statistical model algorithm among this group may not need to establish an a priori hypothesis of the target's maneuverability, and it has a good performance in tracking when the target is non-maneuvering.