非线性系统的神经网络自适应逆控制

Neural Network Adaptive Inversion Control for Nonlinear Systems

提出了非线性系统的神经网络自适应逆控制方法。设计中使用了 2个神经网络 ,经离线训练的NN1实现非线性系统的逆 ,在线网络NN2用于补偿逆误差和系统的动态特性变化。对一非线性系统的仿真结果表明 ,神经网络自适应逆控制能够提高系统的动态性能 ,并且具有较好的鲁棒性。

A method of neural network adaptive inversion control for nonlinear systems is proposed.Two neural networks are used,the inversion of nonlinear systems is realized by off line trained NN1,on line NN2 is used to compensate inversion error and varieties of system dynamic property.Simulation results of a nonlinear system indicate that neural network adaptive inversion control can improve system dynamic property and has better robustnees.