摘要
许多系统所固有的不确定性和非线性,难以建立确切的数学模型,其逆模型也难以建立,因此自适应逆控制的应用受到了限制。针对这种情况,利用神经网络可以逼近任意非线性函数的能力来构建系统的逆,用来作为控制器;另一个神经网络作为被控对象的辨识模型来构建一个控制系统。仿真实验结果表明,所设计的控制结构是合理的和有效的,并且具有很强的鲁棒性。
It’s difficult to erect precise mathematical model for many systems because of their proper uncertainty and non-linearity, so their inverse model does, and the application of Adaptive Inverse Control(AIC) system is limited. In allusion to such case, we can use the Artificial Neural Network (ANN) that has the ability to gain on any nonlinear functions to erect a inverse model, and make the ANN as a controller, and the other ANN as a identification model to form a controlled system. Through the simulated experiment, it approved that the system designed is reasonable and effective, in addition, it possesses very strong robustness.
出处
《电气技术》
2007年第7期62-66,共5页
Electrical Engineering