基于DRNN在线辨识的真空炉温度控制

Single Neuron PID Control for Vacuum Furnace’s Temperature Based On Diagonal Recurrent Neural Network On- Line Identification

提出一种基于DRNN在线辨识的单神经元PID控制方案,由DRNN对系统进行在线辨识,建立其在线参考模型并为单神经元控制器提供梯度信息,从而实现控制器参数的在线调整。仿真结果表明,对象参数非线性时变时,该控制方法温度控制精度高,动态特性好,收到了良好的效果。

This paper puts forward a single neuron PID controlling project based diagonal recurrent neural network. It realities on-line identification about the system with DRNN, then it can set up on-line reference model and it also can offer gradient information for single neuron PID controller. As a result it realities on-line adjusting of the controller’s parameter. The simulating results indicate that this controlling way has higher precision and better dynamic characteristic, especially for nonlinear time-variant plants.