摘要
该文提出了一种新的用于非线性非高斯系统状态估计的粒子滤波算法。首先通过基于数值积分的差商滤波器产生重要密度函数,由于这些重要密度函数结合了最新的观测数据,这样采样得到的样本更接近于系统状态的真实后验概率,因此其性能优于标准的粒子滤波算法。最后给出了理论分析和仿真结果,验证了该算法的有效性。
This paper introduces a new particle filter for nonlinear and non-Gaussian systems.The divided difference filter based on numerical integration is used for generating the importance density functions.As it integrates the new observations into system state transition density, which approximates to the state posterior density, the proposed particle filter has the better performance than the conventional one. Finally, the validity of this method is well verified by the computer simulations.
出处
《电子与信息学报》
EI
CSCD
北大核心
2007年第6期1369-1372,共4页
Journal of Electronics & Information Technology
基金
国家部级基金资助课题
关键词
数值积分
差商滤波器
粒子滤波
贝叶斯滤波
目标跟踪
Numerical integration
Divided difference filter
Particle filtering
Bayesian filtering
Target tracking
