电液伺服系统的模糊径向基函数网络监督控制  

Supervised Control Using Fuzzy Radial Basis Function Networks for an Electro-hydraulic Servo System

作　　者：党开放[1] 赵弘[1] 林廷圻[1] 

机构地区：[1]西安交通大学机械电子工程系,陕西西安710049

出　　处：《系统仿真学报》2004年第3期593-596,共4页Journal of System Simulation

摘　　要：针对电液伺服系统存在未知干扰力及参数时变等问题,提出一种新型的模糊径向基函数(简称RBF)神经网络的在线控制方法。该方法基于RBF网络与模糊推理系统的等价性,将模糊推理的思想引入RBF神经网络,从而为RBF隐层数目和中心位置初始化找到一种有效的途径,并采用Kohonen竞争学习机制在线调节网络参数,构成模糊自组织RBF监督控制策略。对典型电液位置伺服系统的仿真结果表明,该方法实时性强,具有良好的鲁棒性和跟踪性能。An online control strategy based on fuzzy radial basis function network is proposed for the tracking control problem of the electro-hydraulic position servo system. Due to the function equivalence between RBF networks and fuzzy inference system, an effective method, which adds fuzzy inference to RBF networks to initialize the number and centers of the receptive fields, is established to avoid the convergence rate slowing that uses random initial values of number and centers of the receptive fields. The Kohonen learning algorithm is used to modify these two kinds of parameters on-line. Supervised control strategy based on fuzzified RBF networks is adopted. The simulation results of a real electro-hydraulic position servo system show the strong robustness and good tracking performance of this control method.

关 键 词：径向基函数网络 模糊推理系统 监督控制 电液位置伺服系统 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]