摘要
为了实现磁悬浮球系统高精度位置控制,提出一种基于模糊神经网络补偿PID控制的磁悬浮球系统位置控制新方法,该控制系统由模糊神经网络辨识器、PID控制器和模糊神经网络控制器组成。模糊神经网络辨识器基于PID控制器所提供的训练数据,建立控制系统误差与控制量之间的动态模型并将网络辨识参数实时传递至模糊神经网络控制器,模糊神经网络控制器基于实时辨识模型计算得到当前周期的补偿控制量,实现对PID控制的在线动态补偿,避免了离线训练过程,且无需建立精确的数学模型。方波信号仿真和实验结果表明:模糊神经网络补偿控制精度分别由PID控制的0.014 2 mm和0.221 1 mm提升至0.006 8 mm和0.073 9 mm,控制系统具有良好动态性能。
Aiming at high accuracy position control problem of magnetic levitation ball system, a new magnetic levitation ball position control method is proposed based on fuzzy neural network compensating PID control in this paper, the control system is composed of fuzzy neural network identifier, PID controller and fuzzy neural network controller. Based on training data provided from PID controller, the dynamic model between the system error and control quantity is established by fuzzy neural network identifier. Meanwhile, identified network parameters are transmitted to fuzzy neural network controller which calculates to be a compensation control quantity currently based on real-time identified model, which realizes online dynamic compensation for PID control, avoids offline training and not to establish a precise mathematical model. The simulation and experiment results of square signal indicate that the fuzzy neural network compensation control achieves a control precision of 0.006 8 mm and 0.073 9 mm, which is better than the control precision of 0.014 2 mm and 0.221 1 mm for PID control. The control system shows excellent dynamic performance.
作者
沈昕璐
莫鑫
周亚南
Shen Xinlu;Mo Xin;Zhou Yanan(School of Mechanical Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China;School of Publishing and Art Design, University of Shanghai for Science and Technology, Shanghai 200093, China)
出处
《农业装备与车辆工程》
2018年第5期59-64,共6页
Agricultural Equipment & Vehicle Engineering
关键词
模糊神经网络
PID
磁悬浮球系统
位置控制
硬件在环控制实验
fuzzy neural network
PID
magnetic levitation ball system
position control
hardware-in-loop control experiment
