lmproved JPDA Algorithm with Measurements Adaptively Censored
Xiangyang Liu, Keke Wang, Peisheng Zhu, Keli Yang · 2012
For the problem of tracking multiple targets in dense clutter with missed detections, the JPDA approach has shown to tend to coalesce neighboring tracks. To improve this situation, the paper proposed an improved approach to adaptively censor the measurements for the state updating by setting a censoring threshold. Monte Carlo simulations with Matlab show that the method is an effective way to avoid track coalescence. On the crossing point of two targets, the position RMS error of the proposed filter appeared to outperform that of the Scaled JPDA proposed in literature. On the remainder of the simulation time, they appeared to perform similarly. The proposed method, therefore, is capable of avoiding track coalescence with less position RMS error.