动态误差时间序列小波神经网络预测模型  

A prediction model based on wavelet neural network for the time series of dynamic errors

作　　者：丁晓牧[1] 金施群[1] 费业泰[1] 

机构地区：[1]合肥工业大学仪器仪表学院,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2003年第6期1127-1130,共4页Journal of Hefei University of Technology：Natural Science

基　　金：十五"国防军工计量重点资助项目(60104208)

摘　　要：基于现代误差修正技术,研究小波神经网络建立的动态测量误差预测模型,以进行误差修正,提高动态测量精度,避免了传统神经网络需要人为干预网络结构参数的不足。文章介绍了建模方法,重点对大轴圆度误差测量过程中的动态测量数据进行实例分析,结果表明,该模型预测精度高,具有重要的应用价值。A prediction model based on wavelet neural network for the time series of dynamic errors is established to correct the measurement errors and enhance the dynamic measuring accuracy based on the modern errors correction technique. The wavelet neural network is used as a substitute for the traditional neural network to avoid the limitations that the structure parameters of the network need to be changed artificaially. In this paper, the modelling method is introduced and the dynamically measured data of the roundness of a large shaft are modelled. The analysis result shows that the prediction accuracy of the wavelet neural network model is high, and the model is very useful in errors correction.

关 键 词：动态误差时间序列 小波神经网络 误差修正 傅里叶变换 小波分析 测量系统 

分 类 号：O241.1[理学—计算数学]