The Method of Neural Network PID Control with Prediction Control

摘要：

This paper presents two improved training method for designing the weights of neural network PID controller, which use Clarke index function and multi-step quadratic index function to train the weights of the neural network PID controller. Therefore it makes the character of the control system better. The simulation results demonstrate the effectiveness of this method.

关键词： Neural network PID control Clarke’s index function Multi-step quadratic index function Train

作者: Shengqi SUN Xuebin LI Liang DONG Shusheng GU

作者单位: Key laboratory of manufacturing industrial integrated automation, Shenyang University, Shenyang, 110 Shenyang Liming Aero-Enging Group Corporation LTD., Shenyang, 110043, China Key laboratory of manufacturing industrial integrated automation, Shenyang University, Shenyang, 110

会议类型: 国际会议

会议名称: The 22nd China Control and Decision Conference(2010年中国控制与决策会议)

会议地点: 徐州

会议语种:英文

页码: 2515-2518

在线出版日期: 2010-05-26（万方平台首次上网日期，不代表论文的发表时间）