摘要
设计了一种基于嵌入式系统的神经网络PID控制器,以ARM芯片为控制器核心,实现对难以建立精确数学模型的非线性系统的自适应控制;控制器采用RBF神经网络对被控对象进行在线辨识,并根据辨识结果对控制器的参数进行在线修正,实现PID控制器的自适应;该控制器体积小、适应能力强且省电;实验结果表明,该控制器可靠性高,响应快,可以在无法确定被控系统数学模型的情况下达到理想的控制效果。
A neural network PID controller based on embedded system is designed,it uses ARM chip as the main controller,and adaptively controls the nonlinearity systems whose mathematic model is difficult to establish.it is using RBF neural netwok to identify the controlled object,and adjust the controller＇s parameters by the identified results,then come into being the adaptive PID controller.the controller is small,energy-saving and stronger adaptive ability.the result of tests shows that this controller has the fast control speed and higher reliability,and it can control the object in effect even the object＇s mathematic model is unknown.
出处
《计算机测量与控制》
CSCD
北大核心
2010年第9期2066-2069,共4页
Computer Measurement &Control
基金
上海市研究生创新基金项目(JWCXSL0802)
