Title :
Improved Quasi-Monte-Carlo particle filter algorithm Based on GRNN
Author_Institution :
Dept. of Commun. & Eng., Eng. Univ. of China Armed Police Force, Xi´an, China
Abstract :
In order overcome computational complexity and low-precise in the Quasi-Monte-Carlo particle filter (QMC-PF), in this paper we propose a new Quasi-Monte-Carlo Particle Filter Algorithm Based on GRNN (NQMC-PF). In the filter, the particles with heavy weight, which is regarded as father-sample, can be transformed into low-disrepancy particles by using QMC. It ensures the diversity of samples. Meanwhile, the weight of offspring-particle is estimated by GRNN. Then the method could add the diversity of samples, improve the precision and speed. Simulation results show that the method could add the diversity of samples, improve the precision and speed. So it has a high application value and the computational complexity is lower.
Keywords :
Monte Carlo methods; particle filtering (numerical methods); recurrent neural nets; GRNN; QMC-PF; computational complexity; improved quasi-Monte-Carlo particle filter algorithm; low-disrepancy particles; offspring-particle; Low-Discrepancy; Neural Network; Particle filter; Quasi-Monte-Carlo;
Conference_Titel :
Communication Technology (ICCT), 2012 IEEE 14th International Conference on
Conference_Location :
Chengdu
Print_ISBN :
978-1-4673-2100-6
DOI :
10.1109/ICCT.2012.6511225