Approximate Bayes multi‐target tracking smoother. Issue 3 (24th January 2019)

• Record Type: Journal Article
• Authors: Li, Quan‐rui; Qi, Bin; Liang, Guo‐long
• Is Part Of: IET radar, sonar & navigation. Volume 13:Issue 3(2019)
• Page Start: 428
• Page End: 437
• Publication Date: 2019-01-24
• Subjects: Gaussian processes -- tracking filters -- target tracking -- set theory -- Bayes methods -- filtering theory -- smoothing methods -- mixture models; labelled random finite set -- Gaussian mixture implementation -- exponential complexity -- m‐best S‐D assignment algorithm -- multiBernoulli filter -- MTT scenarios -- efficient approximation -- filtering density -- smoothing density -- backward smoothing stage -- multitarget transition kernel -- forward–backward smoothing scheme -- approximate Bayes multitarget tracking smoother; Signal processing -- Periodicals; Radar -- Periodicals; Sonar -- Periodicals; Electronics in navigation -- Periodicals; Navigation -- Periodicals; 621.3848
• DOI: 10.1049/iet-rsn.2018.5297
• Languages: English
• ISSNs: 1751-8784

Abstract

Abstract : An approximate Bayes multi‐target tracking (MTT) smoother that is directly applicable to MTT is proposed in this study. By using labelled random finite set, the proposed smoother inherently produced trajectories of the targets. The smoother is based on Bayes forward–backward smoothing scheme, which involves forward filtering followed by backward smoothing. It is shown that if the multi‐target transition kernel involved in the backward smoothing stage considers no birth and death of targets, then the smoothing density is of the same form as the filtering density. To avoid the exponential complexity of computation, an efficient approximation that utilises the m‐best S‐D assignment algorithm to truncate the smoothing density is presented. A Gaussian mixture implementation of the smoother for a linear system is also given. Simulation results in MTT scenarios demonstrate that the smoother outperforms generalised labelled multi‐Bernoulli filter in terms of optimal sub‐pattern assignment metric.