电液位置伺服系统的模糊神经网络控制  

Fuzzy neural network control of electro-hydraulic position servo system

作　　者：仲伟峰[1] 何小溪[1] 

机构地区：[1]哈尔滨理工大学自动化学院,黑龙江哈尔滨150080

出　　处：《电机与控制学报》2008年第4期478-482,共5页Electric Machines and Control

摘　　要：针对电液位置伺服系统中的非线性、参数时变性等复杂因素,设计了一种模糊神经网络控制方案。由于常规的模糊神经网络学习算法具有权值调整复杂,收敛速度慢的缺点,因此采用模糊逐级误差逼近方法来调整模糊神经网络的权值。该算法易于实现,网络权值在线学习速度较快,而且计算量小于其他的常规神经网络学习算法。将该方法应用于电液位置伺服控制系统中,在对象参数摄动情况下,进行了仿真研究。仿真结果表明,采用该方法所设计的控制器满足系统对快速性和稳态精确度的要求,系统的鲁棒性增强,验证了方法的有效性。Aiming at the nonlinear and time-varying parameters characters of electro - hydraulic position servo system, a fuzzy neural network control method was proposed. Owing to the complex weights adjustment and low convergence speed of routine fuzzy neural network learning algorithms, a fuzzy hierarchy er- ror approach （FHEA） was adopted to adjust the weights of fuzzy neural network. This algorithm has fas- ter online network weights learning speed. It is simple to implement and it does not require as many cal- culations as some other classic neural network learning algorithms. Considering the perturbations of object parameters, FHEA was employed in the simulation of electro-hydraulic position servo control system. Simulation results demonstrate the effectiveness of proposed method. The designed controller satisfies the rapid and steady-state accuracy demands, and the robustness of system is increased.

关 键 词：液压伺服系统 模糊神经网络 逐级逼近 学习算法 

分 类 号：TP273.4[自动化与计算机技术—检测技术与自动化装置]