摘要
在系统阐述贝叶斯估计理论的基础上,分析和总结了扩展卡尔曼滤波、Sigma点卡尔曼滤波以及粒子滤波等方法的特点、使用条件以及局限性。介绍了扩展卡尔曼滤波方法并指出其缺陷;介绍了无轨迹卡尔曼滤波、中心差分卡尔曼滤波等确定性采样方法,并在加权统计线性回归意义下将其归结为Sigma点卡尔曼滤波方法;介绍了随机采样方法—粒子滤波方法,并指出其主要的研究方向。最后,对非线性滤波的发展趋势进行了展望。
The Bayesian estimation theory is elaborated in detail. The characteristics, utilization and limitation of various estimation algorithms, such as extended Kalman filter, Sigma point Kalman filter, and particle filter, etc, are analyzed and summarized. The extended Kalman filter and its flaw are presented. The deterministic sampling methods, unscented Kalman filter and central difference Kalman filter, are introduced and unified within weighed statistical linear regression framework. The random sampling method, namely, particle filter, as well as its main research direction is introduced. Finally, further development trends of nonlinear filter are pointed out.
出处
《指挥控制与仿真》
2009年第5期1-5,共5页
Command Control & Simulation
关键词
贝叶斯估计
卡尔曼滤波
加权统计线性回归
粒子滤波
Bayesian estimation
Kalman filter
weighed statistical linear regression
particle filter