论文部分内容阅读
提出了一种对非线性系统的神经网络自学习控制方法．基于逆动力学控制的思想，构造了神经网络结构一致的控制器和辨识器．辨识器采用多层前向网络结构和广义Ｄｅｌｔａ学习规则算法实现了对系统逆动力学模型的动态辨识，并通过在线动态传递权值给神经网络控制器的方法实现了神经网络辨识器和神经网络控制器的有机结合，从而使整个控制系统具有很强的自适应和自学习能力．所提出的控制方案可适用于不含滞后环节和包含滞后环节的非线性系统．仿真结果证明了这种控制方法的有效性．
A neural network self-learning control method for nonlinear systems is proposed. Based on the idea of inverse dynamics control, a controller and a recognizer with the same structure of neural network are constructed. The recognizer uses multi-layer forward network structure and generalized Delta learning rule algorithm to realize the dynamic identification of the inverse kinematics model of the system, and realizes the neural network identifier and neural network The organic combination of the controller, so that the entire control system has a strong self-adaptive and self-learning ability. The proposed control scheme can be applied to nonlinear systems without hysteresis and hysteresis. Simulation results show the effectiveness of this control method.