Title :
Robust track fusion
Author :
Black, J.V. ; Reed, C.M.
Author_Institution :
DERA, Malvern, UK
fDate :
Aug. 31 1999-Sept. 3 1999
Abstract :
When local tracks are fused to produce a global track, allowing for dependence, the approximations necessary for economic track representation can result in an unlocalised, posterior track distribution and tracker failure. This paper describes two approaches to robust track fusion. Each produces a modified version of the usual formula which undoes the adverse effects of the approximations. The general principles are applied here to the case of Gaussian mixture modelling of tracks.
Keywords :
Gaussian processes; approximation theory; mixture models; sensor fusion; Gaussian mixture modelling; approximation effects; robust track fusion; Approximation methods; Bayes methods; Economics; Probability density function; Robustness; Sensors; Target tracking; architectures; robust track fusion;
Conference_Titel :
Control Conference (ECC), 1999 European
Conference_Location :
Karlsruhe
Print_ISBN :
978-3-9524173-5-5
