基于RBF神经网络的轮胎滚动阻力建模研究  

Research on Tire Rolling Resistance Modeling Based on RBF Neural Network

作　　者：毛鑫昕 毛建清[2] 王东哲 MAO Xinxin;MAO Jianqing;WANG Dongzhe(Dalian Maritime University,Dalian 116026,China;Zhongce Rubber Co.,Ltd,Hangzhou 310008,China)

机构地区：[1]大连海事大学理学院,辽宁大连116026 [2]中策橡胶集团有限公司,浙江杭州310008

出　　处：《橡胶工业》2019年第10期739-743,共5页China Rubber Industry

基　　金：国家自然科学基金资助项目(61671099)

摘　　要：建立径向基函数(RBF)神经网络轮胎滚动阻力模型,充分利用RBF神经网络模型逼近精度高、训练速度快、无局部极小等优点,对轮胎滚动阻力进行全面、准确的预测。结果表明,轮胎滚动阻力RBF与反向传播算法(BP)神经网络模型预测值的平均相对误差分别为2%和6%左右,RBF神经网络模型在训练和预测结果上均有更大优势,能够有效预测轮胎滚动阻力。A tire rolling resistance model based on radial basis function(RBF)neural network was established,taking full use of the advantages of RBF network model,such as high approximation accuracy,fast training speed and no local minimum,to predict tire rolling resistance comprehensively and accurately.The results showed that the average relative errors of the prediction values of tire rolling resistance RBF neural network model and back propagation(BP)neural network model were about 2%and 6%,respectively.RBF neural network model showed advantages in training and prediction results,and could effectively predict tire rolling resistance.

关 键 词：轮胎 滚动阻力 模型 径向基函数 反向传播算法 神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TQ336.[自动化与计算机技术—控制科学与工程]