Data association for PHD filter based on MHT

Author

Wang, Yang ; Jing, Zhongliang ; Hu, Shiqiang

Author_Institution

Sch. of Electron., Shanghai Jiao Tong Univ., Shanghai

fYear

2008

fDate

June 30 2008-July 3 2008

Firstpage

1

Lastpage

8

Abstract

The main drawback of probability hypothesis density (PHD) filter is that it canpsilat identify the trajectories of the different targets. Data association for PHD filter based on multiple hypotheses tracking (MHT) is presented to solve the problem. The track-oriented MHT is used to perform data association on the output of PHD filter. An adaptive Kalman filter based on ldquocurrentrdquo statistic model, combined with MHT, is implemented to track maneuvering targets. Two examples are given to test the performance of the new method. Monte Carlo simulation results show that this approach is computationally feasible and effective for associating multi-targets in dense clutter environments.

Keywords

Kalman filters; Monte Carlo methods; adaptive filters; probability; sensor fusion; statistical analysis; target tracking; Monte Carlo simulation; adaptive Kalman filter; data association; maneuvering target tracking; multiple hypotheses tracking; probability hypothesis density filter; statistic model; data association; probability hypothesis density; track-oriented MHT;

fLanguage

English

Publisher

ieee

Conference_Titel

Information Fusion, 2008 11th International Conference on

Conference_Location

Cologne

Print_ISBN

978-3-8007-3092-6

Electronic_ISBN

978-3-00-024883-2

Type

conf

Filename

4632445