基于神经网络的Smith补偿PID控制设计

针对工程实践中常见的纯滞后对象，在Ｓｍｉｔｈ补偿控制基础之上，利用人工神经网络的非线性映射功能对控制对象进行在线辨识，达到对时滞补偿预报的目的；利用神经网络ＰＩＤ控制器（Ａｄａｌｉｎｅ网络）代替常规控制器，实现了对时滞复杂对象的在线自适应控制；并根据ＩＴＡＥ性能指标原则对神经网络控制器参数进行整定，得到一组经验公式。仿真结果验证了本文神经网络控制方案的有效性。

Design of Smith Compensation-PID Control Based on Neural Networks

Based on the Smith compensation control theory,the prediction of time lag compensation is designed by use of the nonlinear mapping function of neural network on line identification,and then the conventional PID controller is replaced with NN PID controller(the Adaline networks are used)to take the adaptive control for the time lag object.And a group of empirical formulations are obtained through determination of the parameter of NN controller according to ITAE criterion.The result of simulation shows that the present scheme is very effective.