Implementing Flatness Recognition Based on GA-CRBF Network by DSP
ZHANG Xiu-ling;CHENG Yan-tao;QI Qing;HOU Dai-biao;Key Laboratory of Industrial Computer Control Engineering of Hebei Province,Yanshan University;National Engineering Research Center for Equipment and Technology of Cold Strip Rolling,Yanshan University;
In view of poor anti-interference ability,limited recognition accuracy and poor capacity for dealing with uncertainties of the conventional RBF neural network,a cloud model was introduced into the Network to establish a novel flatness recognition model. MATALB simulation results showed that new GA-CRBF network could correctly identify flatness defect,with the recognition accuracy and anti-interference capacity increased by 73% and 83% respectively,compared with the traditional RBF network. With GA-CRBF network written into the DSP chip,it can correctly identify defective flatness after running,verifying the feasibility of its application into engineering,which lays foundation for an application of neural network into practical engineering.