摘要
对具有不确定性和干扰项的重复非线性注塑机控制系统,尤其是注射速度的控制,提出基于神经网络的迭代学习控制器,其中迭代学习控制器设计为神经网络控制器,它以前馈方式作用于对象。PD反馈控制器用于使系统达到稳定,同时和前馈的神经网络学习控制器一起使系统达到理想的控制效果。仿真结果表明,该控制器可以随着迭代次数的增加有效减小跟踪误差。
Iterative learning controller based on neural network is presented for repetitive nonlinear plastic injection molding machine with uncertainty and disturbance, especially for injection speed control. The iterative learning controller is designed as the neural network controller, which acts on objects in the form of feed forward. By PD controller which is used to make system more stable, the system can reach ideal control effects with the feed forward neural learning controller. The simulation results indicate that the controller can effectively reduce tracking errors along with the increasing iterative times.
出处
《计算机辅助工程》
2005年第4期71-74,共4页
Computer Aided Engineering
关键词
神经网络
迭代学习
反馈控制
逆模型
PID
neural network
iterative learning
feedback control
inverse dynamic model
PID
