摘要
首先分析了基于贝叶斯理论的粒子滤波算法的原理;然后在分析采样-重要性-重采样算法基础上讨论了粒子滤波算法存在的主要问题,研究了一种使用UKF产生重点密度函数的粒子滤波算法(UPF);最后通过实例将该算法与粒子滤波算法进行比较,仿真结果表明UPF算法运算时间低于粒子滤波算法。
The Particle Filter (PF) algorithm based on Bayesian theory is analyzed. Then, based on the analysis of a standard algorithm of sampling-importance-resampling filter, the problems of PF are discussed and another PF algorithm which importance density is generated by UKF is researched. Finally, the comparison of two algorithms" performance is presented by an application example, the simulation results have shown that calculation time of the unscented particle filter (UPF) is less than PF.
出处
《微电子学与计算机》
CSCD
北大核心
2006年第11期41-43,共3页
Microelectronics & Computer
关键词
粒子滤波
贝叶斯估计
非线性滤波
UPF
Particle filter, Bayesian estimation, Nonlinear filter, Unscented particle filter