机构地区: 北京航空航天大学能源与动力工程学院
出 处: 《系统工程与电子技术》 2006年第9期1391-1393,共3页
摘 要: 通过神经网络技术对柔性机构复杂的非线性动态响应进行辨识,建立了柔性机构运动参数的辨识模型。利用径向基函数(radial basis function,RBF)神经网络优异的非线性逼近能力,建立了柔性机构动态响应的辨识模型。将机构的驱动力矩、阻尼力矩和非线性运动参数分别作为RBF神经网络的输入样本和期望输出样本。建立了RBF神经网络的拓扑结构,利用样本数据对其进行训练。通过空间站柔性展开机构模型进行动态响应的辨识,结果表明辨识的收敛速度快,精度高。该方法为复杂大系统的建模和分析提供了一种理想的途径。 The aim of the research is to setup kinematical parameters identification model of flexible mechanism. Taking advantage of radial basis function (RBF) artificial neural network, the model is realized to identify complicated nonlinear dynamic response of flexible mechanism. Driving moments, damps and nonlinear dynamical parameters are considered as the inputs and outputs (samples) of RBF that constructed to be trained. A flexible expand mechanism of space station is employed to test this identification method. The simulation results indicate that it is a better approach with very well convergence properties and higher fidelity. This method provides an available way of model identification for complex large system.
关 键 词: 系统辨识 径向基函数 柔性机构 非线性 动态响应
领 域: [自动化与计算机技术] [自动化与计算机技术]