不确定多时滞系统动态自适应神经网络控制

Dynamic adaptive neural network control of systems with uncertain multiple-time-delays

针对一类具有不确定多时滞的非线性系统,提出一种由时滞补偿器和动态结构自适应神经网络所构成的控制器。通过设计时滞补偿器中的参数自适应调节规律来消除多时滞对控制输入的影响,再引入动态自适应神经网络,利用其隐层神经元个数可以随着逼近误差的增大而在线增加的特点,获得满意的逼近精度,提高控制性能。最后,对时滞混沌系统进行仿真,表明该方法的有效性。

For a class of nonlinear systems with uncertain multiple time delays, a kind of control method of combining a kind of time-delay compensator and a dynamic adaptive neural network is presented. In the way of regulating parameters adaptively to eliminate the influence on the input caused by the uncertain multiple time- delays. While the dynamic adaptive neural network adopted has the alterable hidden units which increase by the approximation error growing for the better precision. The given demonstration shows the presented control method is effective in a chaotic system with uncertain multiple time-delays.