摘要
对工业控制领域中非线性系统,采用传统PID控制不能获得满意的控制效果。采用基于梯度下降算法优化RBF神经网络,构建其模型,进而编写M语言程序。以整定PID控制器的参数,使系统输出近似跟踪输入。该方法只需给出粗略的PID控制参数,系统的性能依靠神经网络寻优调整,从而可有效地解决经典PID控制方法中控制参数整定困难的问题,且可克服由PID控制参数整定不准给系统带来的不良影响。
Since it can＇t acquire satisfied control result industry control field. Radial basis function neural network is by using traditional PID control for non-linear systems in optimized based on gradient descent algorithm, and its model is constructed, then the M language program is written. By adjusting parameters of PID controller, the output approximately tracks the input. The method needs only initial parameters of PID controller, and the system performance relies on optimization adjustment of neural network. Thus, it can effectively solve parameter adjustment difficulties in classic PID control method, and can overcome adverse effects created by inaccurate parameter adjustment of PID control.
出处
《兵工自动化》
2009年第1期45-46,50,共3页
Ordnance Industry Automation
关键词
PID
RBF神经网络
参数整定
PID
Radial basis function neural network
Parameter adjustment
