摘要
根据实际生产现场对控制的要求,提出了一种将PID控制与神经网络相结合,采用BP网络辨识未知的被控对象,使用单纯形算法寻找最优的PID控制参数,控制与学习并行的控制方案.并对二阶对象进行仿真研究,将其与单位阶跃响应进行了比较.
Combining the PID control with neural network, using the BP algorithm for system identification, and using simplex method to optimize parameters for PID controller, a parallel scheme for controlling between studying and control is brought forward. This new control scheme could adjust the control action according to the changes of the environment. Finally, some simulation results and the comparison to the step responses for the second-order systems are given.
出处
《西南师范大学学报（自然科学版）》
CAS
CSCD
北大核心
2004年第1期41-44,共4页
Journal of Southwest China Normal University(Natural Science Edition)
关键词
BP神经网络
PID控制
系统辨识
单纯形寻优
PID control
neural network
BP neural network
system identification
simplex method
