电液位置伺服系统的模糊CMAC神经网络控制研究

本文提出了一种模糊ＣＭＡＣ神经网络的结构和算法。该网络通过权系数的在线学习,实现修正模糊逻辑。将此网络作为前馈补偿单元,与ＰＤ控制器一志构成一种自学习控制器,并对带有未知负载干扰的某电液位置系统进行动态仿真。结果表明,该控制器具有强的鲁棒性以及良好的跟踪特性。

Fuzzy CMAC neural network control of the electrohydraulic position servo systems subjected to unknown disturbances

A fuzzy cerebellar model articulation controller (FCMAC) is proposed for the tracking control of the nonlinear electrohydraulic position servo systems subjected to unknown external load disturbances. Fuzzy logical rules are improved though on line learning of FCMAC weights. The learning controller is composed of a linear controller and a FCMAC neural network that is used as feedforward compensator. The simulation results demonstrate that the proposed controller is of good robustness and tracking property for the nonlinear electrohydraulic position servo systems subjected to unknown external load disturbances.