基于微分器与神经网络的PID控制算法研究  

Research of PID Control Algorithm Based on Tracking Diferentiator and Neural Network

作　　者：郭士荦 许江宁[1] 何泓洋[1] 李京书[1] 

机构地区：[1]海军工程大学导航工程系,湖北武汉430033

出　　处：《计算机仿真》2014年第9期379-382,401,共5页Computer Simulation

基　　金：国家自然基金(61104184);国家重大科学仪器开发专项(2011yq12004502);教育部重点实验室基金(201001)

摘　　要：在工控系统微分器设计优化问题的研究中,由于在工业生产实践中,传统PID参数整定方法难以得到合理优化的PID参数,传统增量式PID控制存在抗干扰性较差、控制精度不高的问题。针对上述问题,提出了基于跟踪微分器与神经网络的PID控制算法。对跟踪微分器的原理进行了阐述,通过跟踪微分器实现位置信号滤波与速度信号求解;对增量式自适应PID控制算法进行了改进。在上述基础上,构建了基于微分跟踪器与神经网络的PID控制器。对控制算法进行了仿真,仿真结果表明,提出的控制算法具有抗扰动能力强、控制精度高等优点。In the industrial practice, traditional PID parameter setting method is difficult to get reasonable and optimal PID parameters, the traditional incremental PID control algorithm shows a poor performance in anti-interference performance and control accuracy. In allusion to the problems mentioned above, this paper proposes a PID control algorithm based on tracking diferentiator and neural networks, expounds the principle of the tracking differentiator, re- alizes the filtering of the position signal and the calculation of the velocity signal, ameliorates the traditional incremen- tal PID control algorithm. Then on this basis, designs a PID controller based on tracking diferentiator and neural networks. The simulation results indicate that the control algorithm proposed in this paper performs better in anti-inter- ference performance and control accuracy.

关 键 词：控制算法 跟踪微分器 神经网络 参数自整定 

分 类 号：U666.1[交通运输工程—船舶及航道工程]