基于ANN逆模型的非线性系统内模控制结构方案及仿真研究

基于逆模型的非线性系统控制器设计方法已得到了快速发展,并成为设计控制系统的一种重要方法。但对大量复杂、未知和不确定的非线性对象,建立具有一定通用性的逆模型控制器,在实际非线性系统设计中是非常困难的。文中基于系统内模控制（IMC）原理和神经网络（ANN）的非线性映射、大规模并行处理等特性,设计了非线性系统的ANN正模型辨识器及逆模型控制器,基于IMC原理和ANN理论建立了非线性系统的IMC结构方案。大量仿真研究表明,设计的基于ANN逆模型的IMC结构方案是有效的,ANN逆模型控制器具有良好控制性能,且其结构简单,具有一定通用性。

Simulation and Research on IMC Structure Scheme Based on ANN's Inverse Model for Nonlinear System

It was one of important method for controller design of the nonlinear control system based on the system＇s inverse model, which was rapidly got development in recently years. However for massive complex, unknown and the indefinite nonlinear object, in order to model the inverse model with general use, there were many difficulties in the controller design of the actual nonlinear systems. Based on principle of IMC （Internal Model Control） and the good performance of the ANN＇s （Artificial Neural Network） nonlinear mapping and so on, the ANN＇s feedforward model identifier Mf and the ANN＇s inverse model controller Mi were designed in this paper. The IMC structure scheme based on the IMC principle and the ANN＇s theory was designed for nonlinear system. Massive simulation researches show that the IMC structure scheme designed based on the ANN＇s inverse model is successful and effective, and the ANN＇s inverse model controller Mi has the well control performance and general use