一种基于CMAC神经网络的板形模式识别新方法

针对传统的板形识别模型在识别板形时,由于板宽的变化需要不同拓扑结构的神经网络才能完成板形模式识别任务,网络学习工作量大,网络存在收敛速度慢,易陷入局部极小等结构性能不佳的问题,建立了一种新的基于CMAC神经网络的板形模式识别模型。该模型利用欧式距离差得到网络的输入神经元,并在权值更新算法中引入了动态学习率。通过仿真实验表明该方法简单、实用,识别精度较高,克服了传统的识别模型的缺点和不足,有效地提高了板形模式识别模型的速度和精度。

A new method for flatness pattern recognition based on CMAC network

In traditional flatness pattern recognition neural network, the topologie configurations need to be rebuilt with the width change of cold strip, so learning assignment is large, convergence is slow and it is easily ended in local minimal in the network. Moreover, the structure of the traditional neural network has been proved that the model is time-consuming and complex according to the experience. So a new approach of flatness pattern recognition is proposed based on the CMAC neural network. The difference of fuzzy distances between samples and the basic patterns are introduced as the inputs of the CMAC network. Simultaneity adequate learning rate is improved in the error correction algorithm of this neural network. The new approach with the advantages, such as fast learning speed, good generalization, and easiness to implement, is efficient and intelligent. The simulation results show that the speed and accuracy of the flatness pattern recognition model are improved obviously.