Structural identification using neural network and Kalman filter algorithms
Structural identification using neural network and Kalman filter algorithms
复制标题
使用神经网络和卡尔曼滤波器算法进行结构识别
DOI：
10.2208/jscej.1997.563_1
复制
发表时间：
1997
期刊：
影响因子：
--
通讯作者：
Makoto Sato
中科院分区：
文献类型：
--
作者：
Tadanobu Sato;Makoto Sato
识别结构系统的动力特性。相关的神经网络的学习算法的特点进行了讨论的背景下，系统识别。由于神经网络的自学习性质，所识别的动态特性受到示教信号中包含的噪声水平的强烈影响。本文提出了一种利用卡尔曼滤波技术识别结构系统动态特性的方法，该方法能克服示教信号中的噪声干扰。通过对线性和非线性结构系统动力响应特性识别的算例，验证了算法的稳定性和鲁棒性。
The dynamic characteristics of a structural system are identified. The relevant neural network characteristics of a learning algorithm are discussed in the context of system identification. Because of the self-learning nature of the neural network the dynamic characteristics identified are strongly affected by the level of noise contained in the teaching signals. A method to identify the dynamic characteristics of a structural system proof against contaminating noise in teaching signals has been developed with the aid of the Kalman filtering technique. Numerical examples to identify dynamic response characteristics of linear and nonlinear structural systems are worked out to demonstrate the stability and robustness of the proposed algorithm.