多层前馈感知器的高阶序贯非线性Kalman滤波学习算法

邓志东; 孙增圻

引用本文:	邓志东,孙增圻.多层前馈感知器的高阶序贯非线性Kalman滤波学习算法[J].控制理论与应用,1994,11(3):381~384.[点击复制]
	DENG Zhidong and SUN Zengqi.Learning Algorithm of Multilayer Feedforward Perceptron with Higher Order Sequential Nonlinear Kalman Filter[J].Control Theory and Technology,1994,11(3):381~384.[点击复制]

多层前馈感知器的高阶序贯非线性Kalman滤波学习算法

Learning Algorithm of Multilayer Feedforward Perceptron with Higher Order Sequential Nonlinear Kalman Filter

摘要点击 1172 全文点击 532 投稿时间：1992-12-26 修订日期：1993-06-02

查看全文查看/发表评论下载PDF阅读器

DOI编号

1994,11(3):381-384

中文关键词多层前馈感知器学习算法非线性Kalman滤波高阶序贯估计

英文关键词 multilayer feedforward perceptron learning algorithm nonlinear Kalman filtering higher order sequential estimation

基金项目

作者	单位
邓志东,孙增圻	清华大学计算机系

中文摘要

本文提出了高阶序贯非线性增广Kalman滤波（SEKF），并将其应用于多层前馈感知器（MLPs）的学习问题。文中给出了MLPs的SEKF算法，得到了与BP算法类似的正向与q反向传播过程，并且详细地推导了核心的量测Jacobian矩阵。结合一非线性正弦函数，DEKF和SEKF的仿真结果被进一步给出。

英文摘要

In this paper a higher order sequential nonlinear extended Kalman filter （SEKF）, based on the DEKF algorithm given by ref. [4], is proposed and applied to the learning problem of MLPs. The simulation result is shown that SEKF is superior to DEKF in the filtering accuracy and the required learning number except the slightly increased computational complexity.