一种新的复合神经网络控制器

A novel composite artificial neuron net controller

目的介绍一种新的复合神经网络控制器的结构。方法结合传统控制、神经网络监督控制、神经网络逆控制的优点提出复合控制器方案,建立用它控制水箱系统的水位和温度的仿真模型,并进行计算机仿真。结果仿真实验表明,复合神经网络控制器中的PD控制器、神经网络监督控制器、神经网络逆控制器,可以协同工作,也可以在某种程度上单独工作。结论复合神经网络控制系统的鲁棒性和抗干扰性得到提高。

Aim A novel composite artificial neuron net（ANN） controller is introduced. Methods A novel composite controller based on the combination of conventional control, ANN supervised control and ANN inverse control is given; a simulation model controlling water level and temperature is generated, and simulations are performed. Results The simulations demonstrate that the conventional PD control, ANN supervised control and ANN inverse control can cooperate or work separately in some way. Conclusion The composite ANN controller enhances the robustness and anti-interference of the system.