论文部分内容阅读
针对自适应神经网络跟踪控制问题,提出一种确定逼近域的方法.采用参考信号取代未知非线性函数中的系统输出,神经网络用于逼近以参考信号为输入的未知不确定项.可以利用参考信号的界预先确定神经网络逼近域,再采用自适应鲁棒方法处理由于函数输入置换所引起的另一类不确定项.所得到的闭环系统是全局稳定的.仿真实例说明了该控制方法的有效性.
Aiming at the problem of adaptive neural network tracking control, a method to determine the approximation domain is proposed. The reference signal is used to replace the system output in the unknown nonlinear function, and the neural network is used to approximate the unknown uncertain signal with the reference signal as input. The signal boundary pre-determines the approximation domain of the neural network, and then uses the adaptive robust method to deal with another type of uncertainties caused by the input substitution of the function. The closed-loop system is globally stable. The simulation example shows that the control method Effectiveness.