DocumentCode :
2760303
Title :
Evolutionary MCMC Particle Filtering for Target Cluster Tracking
Author :
Carmi, Avishy ; Godsill, Simon J. ; Septier, Francois
Author_Institution :
Dept. of Eng., Univ. of Cambridge, Cambridge
fYear :
2009
fDate :
4-7 Jan. 2009
Firstpage :
262
Lastpage :
267
Abstract :
A new filtering algorithm is presented for tracking multiple clusters of coordinated targets. Based on a Markov chain Monte Carlo sampling mechanization, the new algorithm maintains a discrete approximation of the filtering density of the clusters´ state. The filter´s tracking efficiency is enhanced by incorporating two stages into the basic Metropolis-Hastings sampling scheme: 1) Interaction. Improved moves are generated by exchanging genetic material between samples from different realizations of the same chain, and 2) Optimization. Optimized proposals in terms of likelihood are obtained using a Bayesian extension of the EM algorithm. In addition, a method is devised based on the Akaike information criterion (AIC) for eliminating fictitious clusters that may appear when tracking in a highly cluttered environment. The algorithm´s performance is assessed and demonstrated in a tracking scenario consisting of several hundreds targets which form up to six distinct clusters in a highly cluttered environment.
Keywords :
Bayes methods; Markov processes; Monte Carlo methods; approximation theory; evolutionary computation; expectation-maximisation algorithm; optimisation; particle filtering (numerical methods); pattern clustering; sampling methods; target tracking; Akaike information criterion; Bayesian method; EM algorithm; Markov chain Monte Carlo sampling method; Metropolis-Hastings sampling scheme; discrete approximation; evolutionary MCMC particle filtering; highly cluttered environment; multicluster target tracking; optimization; Approximation algorithms; Bayesian methods; Clustering algorithms; Filtering algorithms; Genetics; Monte Carlo methods; Particle tracking; Proposals; Sampling methods; Target tracking; Evolutionary MCMC; Markov chain Monte Carlo filtering; Multi cluster tracking; Variational Bayesian EM algorithm;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Digital Signal Processing Workshop and 5th IEEE Signal Processing Education Workshop, 2009. DSP/SPE 2009. IEEE 13th
Conference_Location :
Marco Island, FL
Print_ISBN :
978-1-4244-3677-4
Electronic_ISBN :
978-1-4244-3677-4
Type :
conf
DOI :
10.1109/DSP.2009.4785932
Filename :
4785932
Link To Document :
بازگشت