The Probabilistic Multi-Hypothesis Tracker (PMHT) is an emerging tracking algorithm that appears to have the potential to compete with other well-established tracking algorithms. One of the values that the PMHT brings to the tracking problem is its computational efficiency that grows linearly as the number of targets increases, whereas most tracking algorithms increase exponentially as targets increase. Knowing this, how much does this computational efficiency for the PMHT translate into an algorithm speed advantage? The Multi-Hypothesis Tracker (MHT) was first presented in the late 1970s. Since then significant work has been done in order to improve this robust algorithm, and today the MHT is one of the leading tracking algorithms. Taking an efficient coding of the MHT, it is used as a comparison for the PMHT in terms of algorithm speed. In order to make this comparison objectively, the PMHT is run against the MHT in a common environment. Results have been produced for both the single target scenario and the multiple target scenarios.
[1]
R. Streit,et al.
Probabilistic Multi-Hypothesis Tracking
,
1995
.
[2]
Darin T. Dunham,et al.
Hybrid tracking algorithm using MHT and PMHT
,
2002,
SPIE Defense + Commercial Sensing.
[3]
Samuel S. Blackman,et al.
Design and Analysis of Modern Tracking Systems
,
1999
.
[4]
Robert J. Dempster,et al.
Reducing MHT computational requirements through use of cheap JPDA methods
,
2001,
SPIE Optics + Photonics.
[5]
Darin T. Dunham,et al.
Tracking multiple targets in cluttered environments with a probabilistic multihypothesis tracker
,
1997
.