【摘要】 为了克服实际控制系统中存在的非线性和参数时变性所引起的常规控制器控制性能恶化,提出了一种新的控制方法———基于RBF神经网络整定的模糊控制,并在文中给出了具体算法.该控制方法以无量化解析模糊控制为主体,采用RBF神经网络对控制对象进行辨识,然后利用辨识所得到的Jacobian信息在某一给定的控制性能指标下对控制参数进行在线调整.将其应用于二次调节控制系统,并对系统进行了仿真.系统采用不同控制器的仿真曲线表明:基于RBF神经网络整定的模糊控制具有更好的控制性能.更多还原

【Abstract】 In order to overcome the property deterioration of the controller due to the non-linear and the parameter changing character in the actual control system,this paper puts forward a new kind of control method- the fuzzy controller rectified by RBF neuro-network,and provides the concrete algorithm in detail.This algorithm,with the non-quantitative analytic fuzzy controller as its core,can carry out the on-line parameter adjustment according to a certain norm of control property through the following process: Firstly,the controller identifies the control system by the usage of RBF neural network.Secondly,from the system identification the controller can get the Jacobian information of the control system.Thirdly,the controller adjusts and optimizes the parameter of the controller.This paper applies this algorithm in the hydraulic system with the secondary unit,and makes its simulation.The simulation curves indicate that the Fuzzy Controller Rectified by RBF Neuro-network has advantages over other controller in control properties.更多还原