DocumentCode :
262834
Title :
Implementation of the homotopy particle filter in the JPDA and MAP-PF multi-target tracking algorithms
Author :
Bell, Kristine L. ; Stone, Lawrence D.
Author_Institution :
Metron, Inc., Reston, VA, USA
fYear :
2014
fDate :
7-10 July 2014
Firstpage :
1
Lastpage :
8
Abstract :
In a conventional particle filter, the information update step can suffer from particle degeneracy if the likelihood function is concentrated on only a few particles. The homotopy particle flow method has been developed to implement the information update in an entirely different manner by using a particle flow function to migrate the particles to regions of the target state space that provide a good representation of the posterior distribution. In this paper we demonstrate how the homotopy particle filter can be implemented with the JPDA and MAP-PF multi-target tracking algorithms, and compare performance to the conventional resampling method.
Keywords :
maximum likelihood estimation; particle filtering (numerical methods); probability; sensor fusion; target tracking; JPDA; MAP-PF; homotopy particle filter; homotopy particle flow method; information update step; joint probability data association; maximum a posteriori penalty function; multitarget tracking algorithms; posterior distribution; Atmospheric measurements; Covariance matrices; Particle filters; Particle measurements; Probability density function; Target tracking; Vectors; JPDA; MAP-PF; homotopy particle filter; multi-target tracking; particle filter;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Information Fusion (FUSION), 2014 17th International Conference on
Conference_Location :
Salamanca
Type :
conf
Filename :
6916024
Link To Document :
بازگشت