摘要
提出了一种基于混合PSO和RBF神经网络的自整定分数阶PID控制器的设计方法.该控制器主要由三个部分组成:(1)分数阶PID控制器直接控制被控对象;(2)利用细菌觅食算法和粒子群算法混合优化分数阶PID参数值,作为初始值;(3)利用RBF神经网络具有以任意精度逼近非线性函数及训练速度快的优点,在线整定分数阶PID值,并完成对被控对象的Jacobian信息辨识.实验仿真结果表明:该控制器具有响应速度快、收敛精度高、鲁棒性强等特点,可适用于不同的对象和过程,特别是复杂的、无确定数学模型的控制系统.
A self-tuning fractional order PID controller based on hybrid PSO and neural networks is presented. It consists of three parts. In the first part,a fractional order PID controller directly controls the object. In the second part,a group of fractional order PID parameters are obtained by hybrid PSO. The third part,RBF neural networks is used to optimize and adjust the fractional order PID parameters on-line by exploiting the nonlinear mapping capabilities of neural networks and training fast and get Jacobian information for the object. The simulation results show that the controller has a fast response speed,high convergence and robustness. It can be used to control different objects and processes,specially complex,non-determination mathematical model control system.
出处
《微电子学与计算机》
CSCD
北大核心
2010年第5期157-161,166,共6页
Microelectronics & Computer
