机电工程技术2025，Vol.54Issue(7)：28-32,136,6.DOI:10.3969/j.issn.1009-9492.2025.07.006
神经网络算法拟合双电测组合法修正函数
Correction Function Fitted using Neural Network Algorithm for Dual-configuration Method
摘要
Abstract
With the development of modern science and technology,electronic devices are widely used in medical devices,aerospace and other fields.Sheet resistance is one of the parameters for performance monitoring in the production process of devices,among dual-configuration method is a common test method.The correction function of dual-configuration method includes auxiliary function and thickness function,whose original form is difficult to calculate directly.In order to solve the problem,the neural network algorithm is used to fit the correction function.Above all,the iterative algorithm is presented and its limitation is discussed.Although the iterative algorithm can solve the exact value of the correction function,it may take a long time to calculate.Then,a neural network fitting algorithm is proposed.The data set obtained by the iterative algorithm is used as the input of the neural network,so that it can approach the correction function.When the objective function converges to the minimum value,the fitting function of the correction function can be obtained.By analyzing the relative error between the fitting value and the exact value,the results show that the absolute value of the fitting relative error is less than 5×10-5,which shows a high fitting accuracy.These fitting functions can be directly applied to dual-configuration method measuring,which effectively simplifies the data processing of the thickness function and the auxiliary function and brings convenience to the measurement of sheet resistance.关键词
修正函数/双电测组合法/神经网络/函数拟合/范德堡函数Key words
correction function/dual-configuration method/neural network/function fitted/van der Pauw function分类
电子信息工程引用本文复制引用
冯金良,王亚林,王尚前,廖晓娟,王可..神经网络算法拟合双电测组合法修正函数[J].机电工程技术,2025,54(7):28-32,136,6.基金项目
江西省重点研发计划项目(20192BBEL50018) （20192BBEL50018）
