基于遗传神经网络的自适应PID控制器的设计

Design of self-adaptive PID controller based on genetic neural networks

提出了一种基于遗传算法和神经网络的自适应PID控制器的设计方法。该控制器主要由三个部分组成:利用遗传算法优化PID参数,和RBF神经网络结合,对被控对象逼近,搜索出一组准优的初始参数;RBF神经网络完成对被控对象Jacobian信息辨识;基于单神经元的自适应PID控制器,在线调整PID参数,以确保系统的响应具有最优的动态和稳态性能。仿真结果表明,控制器具有响应速度快,稳态精度高等特点,可用于控制不同的对象和过程。

A self-adaptive PID controller based on genetic algorithm and neural networks is presented.It consists of three parts： PID parameters are optimized by the genetic algorithm,and genetic algorithm combined with the RBF neural networks approaches the controlled object,searching for a group of initial parameters;RBF neural networks get Jacobian information;A self-adaptive PID controller based on the single neural network adjusts the PID parameters on line to insure the optimal dynamic and steady response.The simulation resuhs show that the controller has a fast response speed,high steady precision.It can be used in different objects and processes.