摘要
为对复杂非线性系统进行辨识建模和实施有效控制,分析了基于神经网络的非线性系统逆模型的辨识和控制原理,研究了基于神经网络的非线性系统逆模型补偿的复合控制方法。基于复合控制思想,对常规PID控制器+前馈神经网络逆模型补偿的复合控制结构方案进行了仿真。仿真结果表明,基于神经网络的非线性系统逆模型补偿的复合控制结构方案是有效的、相对简单的网络结构,可提高逆模型的泛化能力和非线性系统的控制精度。
In order to identify, model, effectively control to complicated nonlinear systems, the principle of identification and control of the inverse model for nonlinear system based on artificial neural network （ANN）is analyzed. The compound control structure scheme of inverse model offset for nonlinear system based on ANN is studied. Simulation study is done for the compound control scheme of the conventional PID control method with offset of the inverse model of muhilayer feedforward network ANN. The simulation results show that the compound control structure scheme designed is successful and relatively simple network architecture, the generalization ability of the inverse model and the control precision for nonlinear systems can be raised.
出处
《控制工程》
CSCD
2006年第6期533-535,566,共4页
Control Engineering of China
关键词
神经网络
非线性系统
逆模型
复合控制
仿真
neural networks
nonlinear system
inverse model
compound control
simulation
