摘要
分析了粒子滤波的性能关键,提出了一种新的自适应粒子滤波算法.该算法采用一种新提议分布,即将UKF(Unscented Kalman Filter)与自适应强跟踪滤波器(STF)相结合.新提议分布通过UKF构造粒子群,而粒子群中的每个粒子中的每个sigma点用STF来更新,它可以在线调节因子而使得算法自适应.非线性状态估计仿真试验证实了改进的粒子滤波算法的有效性.
This paper analyzes the keys for the performance of particle filter （PF） and presents a new adaptive PF algorithm. The algorithm adopts a new proposal distribution combining the unscented Kalman filter （UKF） with the adaptive strong tracking filter （STF）. The new proposal distribution adopts UKF to produce the particles, in which each sigma point of every particle is updated by STF. Moreover, the added scaling factor can be adjusted on line to make the algorithm adaptive. Simulated experiments of nonlinear state estimation are finally carried out to confirm the validity of the improved PF algorithm.
出处
《华南理工大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2006年第1期57-61,共5页
Journal of South China University of Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(50405017)
关键词
粒子滤波
状态估计
UKF
自适应滤波
强跟踪滤波器
particle filter
state estimation
unscented Kalman filter
adaptive filtering
strong tracking filter
