针对纯滞后系统神经网络Dahlin控制的研究

Research on Neural Network Dahlin Control for Dead Time System

针对传统的基于Dahlin算法的控制器在大惯性、纯滞后、时变性、非线性对象的控制效果不佳,甚至发生不稳定现象的弱点,提出了以CMAC神经网络与Dahlin算法相结合的控制方法。以CMAC神经网络作为一个前馈控制器,实现时滞系统的自适应稳定控制。仿真实验表明,这种复合控制方法保留了Dahlin算法与CMAC神经网络的各自特长,同时具备学习速度快、适应能力强的优点,具有良好的稳定性和控制效果。

Aiming at the disadvantages of conventional controller based on Dahlin algorithm in controlling objects featuring large inertia, dead time, time varying and nonlinear, i. e. , harmful control effect, even unstable, the control method of combining CMAC neural network and Dahlin algorithm is stated. The CMAC neural network works as the feed forward controller to implement self-adaptive and stable control for time delay system. The experimental simulation verifies that the compound control method keeps the advantages of both Dahlin algorithm and CMAC NN, and features fast leaming speed, high capability of adaptation and excellent and stable control effects.