基于GA与CSA-RBF神经网络辨识的自适应PID控制器

提出了一种基于遗传算法(GA)、克隆选择算法(csA)和神经网络的自适应PID控制器的设计方法.该控制器主要由四部分组成:一是利用遗传算法优化PID参数初始值;二是用克隆选择算法对径向基函数(RBF)神经网络参数初始值优化;三是RBF神经网络完成对被控对象Jacobian信息辨识;四是单神经元PID控制器,学习并在线调整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,该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.

Self-adaptive PID Controller Based on GA and CSA-RBF Neural Networks Identification

A self-adaptive PID controller based on genetic algorithm (GA),clonal selection algorithm (CSA) and neural networks was proposed. It consists of four parts. In the first part,a group of PID parameters are obtained by GA. In the second part,radial basis function (RBF) neural networks parameters are optimized by CSA. In the third part,RBF neural networks get Jacobian information for the object. The fourth part is single neuron PID controller,which studies and adjusts the PID parameters on line to insure the optimal dynamic and steady response. The simulation results show that the controller has a fast response speed,high steady precision. It can be used in different objects and processes.