摘要
将预测控制与神经网络逆动态控制相结合,提出了多步预测性能指标函数下的神经网络逆动态控制方法。该方法用多步预测性能指标函数训练神经网络逆动态控制器的权值,使整个系统具有预测控制的特点,有比通常的神经网络逆动态控制快得多的响应速度和更好的响应性能。仿真实验表明了该方法的有效性。
This paper presents a method of neural network inverse dynamic control under the multi-step predictive index function, which combines the idea of predictive control with the neural network inverse control. The method uses multi-step predictive index function to train the weights of neural network inverse control, thereby the system has the character of predictive control. The method has faster dynamic speed than general neural network inverse control, and has better performance of the response. The simulation results demonstrate the effectiveness of this method.
出处
《控制与决策》
EI
CSCD
北大核心
1999年第4期308-313,共6页
Control and Decision
基金
辽宁省自然科学基金
关键词
预测控制
神经网络
逆动脉控制
性能指标函数
predictive cotnrol, neural network inverse dynamic control, multi-step predictive index function, train