A Markov model for initiating tracks with the probabilistic multi-hypothesis tracker

An important problem in multi-target tracking is track initiation and termination. The tracking algorithm aims to discriminate false detections caused by various sources of interference from valid detections caused by targets of interest. This is a problem of model order estimation. One approach to solving this problem with the Probabilistic Data Association Filter has been referred to as target visibility. This paper shows how the target visibility model can be incorporated into the Probabilistic Multi-Hypothesis Tracker to provide integrated initiation and termination.

[1]  D.A. Gray,et al.  A comparison of track initiation methods with the PMHT , 2002, Final Program and Abstracts on Information, Decision and Control.

[2]  Ning Li,et al.  Target perceivability and its applications , 2001, IEEE Trans. Signal Process..

[3]  Samuel J. Davey,et al.  A comparison of track initiation with PDAF and PMHT , 2002, RADAR 2002.

[4]  R. Streit,et al.  Probabilistic Multi-Hypothesis Tracking , 1995 .

[5]  L. Rabiner,et al.  An introduction to hidden Markov models , 1986, IEEE ASSP Magazine.

[6]  X. R. Li,et al.  Tracker design based on target perceivability , 2001 .

[7]  Robin J. Evans,et al.  Integrated probabilistic data association , 1994, IEEE Trans. Autom. Control..

[8]  Robert W. Sittler,et al.  An Optimal Data Association Problem in Surveillance Theory , 1964, IEEE Transactions on Military Electronics.

[9]  S. B. Colegrove.  Advanced Jindalee Tracker: Probabilistic Data Association Multiple Model Initiation Filter , 1999 .