Comparison of PHD based filters for the tracking of 3D aerial and naval scenarios

Author

Pace, Michele

Author_Institution

Inst. de Math. de Bordeaux (1MB), INRIA Bordeaux - Sud Ouest, Bordeaux, France

fYear

2010

fDate

10-14 May 2010

Firstpage

479

Lastpage

484

Abstract

The Probability Hypothesis Density (PHD) filter is applied to realistic three-dimensional aerial and naval scenarios to illustrate its performance in detecting, initiating and terminating tracks in presence of clutter. Radar measurements are available every two seconds. A comparisons between different approximations of the PHD recursion, namely the sequential Monte Carlo and the Gaussian Mixture approximation, is given on different scenarios using the OSPA metric and different levels of clutter.

Keywords

Gaussian processes; Monte Carlo methods; approximation theory; filters; object detection; radar clutter; radar detection; radar tracking; target tracking; 3D aerial tracking; Gaussian Mixture approximation; PHD based filters; naval tracking; probability hypothesis density filter; radar clutter; radar measurements; sequential Monte Carlo simulation; three-dimensional aerial; track detection; Clutter; Equations; Filters; Monte Carlo methods; Noise measurement; Radar detection; Radar tracking; Recursive estimation; State estimation; Target tracking; Air-Sea Simulation; Multi Target Tracking; PHD Filter;

fLanguage

English

Publisher

ieee

Conference_Titel

Radar Conference, 2010 IEEE

Conference_Location

Washington, DC

ISSN

1097-5659

Print_ISBN

978-1-4244-5811-0

Type

conf

DOI

10.1109/RADAR.2010.5494574

Filename

5494574