DocumentCode :
581813
Title :
A novel multi-sensor multiple model particle filter
Author :
Zhen-tao, Hu ; Xian-xing, Liu ; Jie, Li
Author_Institution :
Inst. of Image Process. & Pattern Recognition, Henan Univ., Kaifeng, China
fYear :
2012
fDate :
25-27 July 2012
Firstpage :
1718
Lastpage :
1722
Abstract :
The large amount of calculation always severely restricts the application domain expansion of particle filter, a novel multi-sensor multiple model particle filtering algorithm based on particle weight optimization is proposed. In the multiple model particle filter framework, the optimization method of particle weight is realized by the extraction and utilization of redundancy and complementary information from latest multi-sensor observations. Due to weaken the adverse influence of random observations noise, the stability and reliability is effective improved. The theoretical analysis and experimental results show that the new algorithm can improve the filter precision but also lessens computational burden in nonlinear system estimation with multi-sensor multiple model characteristic.
Keywords :
optimisation; particle filtering (numerical methods); sensors; application domain expansion; complementary information; multisensor observations; nonlinear system estimation; novel multisensor multiple model particle filtering algorithm; particle weight optimization; random observations noise; redundancy information; reliability; stability; Accuracy; Computational modeling; Equations; Estimation; Mathematical model; Optimization; Particle filters; Interacting Multiple Model; Multiple Model Particle Filter; Particle Weight Optimization; Weighting Fusion;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Control Conference (CCC), 2012 31st Chinese
Conference_Location :
Hefei
ISSN :
1934-1768
Print_ISBN :
978-1-4673-2581-3
Type :
conf
Filename :
6390202
Link To Document :
بازگشت