一种基于CMAC神经网络的模糊控制器设计

Fuzzy controller design based on a CMAC neural network

对于存在外在的环境干扰和系统参数时变的非线性系统来说,传统的小脑模型需要重新学习合适的权重参数,这种学习式的设计方法是相当耗时的,为了改善这种情况,本文提出了模糊控制与小脑模型结合的方式,能够有效地对未知的非线性模型系统进行实时控制。通过仿真的对比试验,这种把小脑神经网络与模糊控制结合起来的控制方法,具有两种控制方法的优点。仿真结果表明,FCMAC控制器具有较高的控制精度、良好的自适应特性。

With regard to nonliear system which contain external environmental interference and time-varying system parameters, the traditional CMAC need to re-study the weight of the appropriate parameters, this study design is very time-consuming. In this thesis, a novel approach of compound fuzzy control and cerebellar model articulation controller （CMAC） can solve the tracking problem of a class of nonlinear systems. Through simulation comparative trial, this compound controller possesses the advantages of the two control methods. The simulation results reveal that FCMAC is capable to guaranteethe system stability and error convergence and make good performance.