论文部分内容阅读
构造了一种用于光学透镜磨抛加工控制的神经网络自适应模糊控制器，同时建立了用于动态估计透镜光圈数的磨抛机模型网络．控制器采用梯度下降学习算法，对网络权重和隶属函数参数进行自动调整，从而可以获得最优隶属函数．系统具有较好的自适应学习功能，仿真结果表明控制性能优于普通模糊控制．
A neural network adaptive fuzzy controller for optical lens grinding and polishing control was constructed. At the same time, a model of the grinding and polishing machine model for the dynamic estimation of lens aperture was established. The controller uses gradient descent learning algorithm to automatically adjust network weights and membership functions, so that the optimal membership function can be obtained. The system has a good self-adaptive learning function, the simulation results show that the control performance is better than the ordinary fuzzy control.