DocumentCode
3216445
Title
Multi-sensor Bearing-only Track Initiation
Author
Hongyan Zhu ; Chongzhao Han ; Chen Li
Author_Institution
Sch. of Electron. & Inf. Eng., Xi´an Jiaotong Univ., China
fYear
2006
fDate
7-11 Aug. 2006
Firstpage
438
Lastpage
443
Abstract
Track initiation is an issue of target detection, which makes a statistic decision about target existence and gives the initial state estimations of the targets. A novel approach for track initiation is proposed in this paper, by using of multi-sensor bearing-only measurements. Firstly, to reduce the complexity of track initiation, two levels association preprocessing are made by geometrical means based on the prior information about sensors, and by position estimations based on multi-sensor bearing-only measurements. Secondly, hybrid quasi-Newton algorithm is adopted to solve the nonlinear least square problem in position estimations. Finally, in order to initiate multi-target tracks and eliminate ghosts, the logic-based technique for track initiation is applied to multi-scan measurement combinations. Simulation results show that the developed method performs well for multi-sensor bearing-only track initiation.
Keywords
Monte Carlo methods; Newton method; least squares approximations; object detection; sensor fusion; state estimation; target tracking; association preprocessing; logic-based technique; multiscan measurement; multisensor bearing-only track initiation; nonlinear least square problem; position estimation; quasiNewton algorithm; state estimation; statistic decision; target detection; IEEE catalog; Image processing; Least squares approximation; Maximum likelihood estimation; Object detection; Position measurement; State estimation; Statistics; Target tracking; Testing; Bearing-only; hybrid Quasi-Newton algorithm; logic-based; measurement combination; track initiation;
fLanguage
English
Publisher
ieee
Conference_Titel
Control Conference, 2006. CCC 2006. Chinese
Conference_Location
Harbin
Print_ISBN
7-81077-802-1
Type
conf
DOI
10.1109/CHICC.2006.280590
Filename
4060553
Link To Document