摘要
DD1滤波是一种基于多元插值计算,将函数按多项式近似展开的非线性滤波方法,相对于扩展卡尔曼滤波而言,它不需要对函数进行微分运算,使得计算大为简化,且具有与扩展卡尔曼滤波相当甚至优于它的估计性能,适用范围较广。对DD1滤波及其在非线性系统识别中的应用进行了分析,给出了计算实例,计算结果表明了该方法的有效性和实用性。
DD1 filter is based on polynomial approximations obtained with a multi-dimensional interpolation formula. In contrast to extended Kalman filter(EKF), the filter is simple to implement as no derivatives are needed and can provide a similar or even superior performance. Using DD1 filtering technique, a method of estimating the parameters of nonlinear systems on the basis of input-output data has been suggested in this paper. A numerical simulation show its suitability and effective for the system even in the case of very severe material nonlinearity.
出处
《信息技术》
2004年第1期11-14,50,共5页
Information Technology
关键词
系统识别
非线性滤波
DD1滤波
参数估计
system identification
nonlinear filter
DD1 filter
parameter estimation
