摘要
基于Magill的Kalman滤波器池结构 ,设计了使用遗传算法的自适应Kalman滤波器 ,给出了离线和在线两种实现方案 .离线方案以辨识滤波参数为主要目的 ,进而可以对状态进行较准确的事后估计 ;在线方案以实时地对状态进行估计为目的 .对滤波参数寻优使用具有良好性能的浮点数编码遗传算法 ,该算法与二进制编码遗传算法相比收敛速度更快、搜索全局最优的能力更强 .
Adaptive Kalman filter with genetic algorithm was designed on the configuration of Magill's Kalman filter bank with. Two realization schemes including off-line and on-line given. The off-line scheme aims primarily at identifying filtering parameters, and the postmortem estimation of states can be obtained more accurately while. The on-line scheme aims at real-time estimation of the states. A well-performed float-coded genetic algorithm was used to optimize the filtering parameters. In comparison with the binary-coded genetic algorithm, this algorithom has higher convergence speed and greater ability of finding global optimum. The efficiency of the proposed adaptive Kalman filter has been verified through simulation.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2003年第6期655-659,共5页
Journal of Harbin Institute of Technology
基金
国家自然科学基金资助项目 ( 60 10 40 0 3 )