摘要
本文利用神经网络的学习能力和非线性映射能力研究了电液加载系统的神经网络直接自适应输出跟踪控制方法，控制器是由一个具有反馈动力学的多层前馈神经网络及其学习算法组成。该控制器不需要被控对象的先验知识，也不依赖于被控对象的辨识模型，能快速跟踪对象的动力学行为，具有良好的自适应性和动态输出跟踪响应性能。
A neural network based direct adaptive output tracking controller is studied for electrohydraulic loading system by using NNs′ learming ability and nonlimear mapping ability in this paper. The controller is composed of a multilayer feedforward neural network with feedback dynamics involved and PID-BP algorithm as its weights learming algorithm. It can qnickly track the system dynamics independently from the identified model and fore information of the controlled system.Simulation results show it has good adaptability and tracking peformances.
出处
《机床与液压》
北大核心
1997年第2期20-23,2,共4页
Machine Tool & Hydraulics
基金
国家重点实验室建设项目
关键词
神经网络
跟踪控制
电液伺服系统
伺服加载
Neural network On-line learning Direct out-Put tracking control Structural loading system
