摘要
针对利用光纤多峰布里渊散射谱解决光纤传感技术中温度和应变同时检测不能获得精确的拟合曲线的问题,将基于K-means聚类方法的RBF神经网络应用到光纤多峰布里渊散射谱的数据拟合中.首先论述了径向基(RBF)神经网络和基于K-means聚类方法的理论知识;其次利用RBF神经网络算法进行数据拟合,得出不同扩散速度影响数据的拟合精度,但是拟合曲线的光滑度和精度不能同时得到保证;最后,采用基于K-means聚类方法的RBF神经网络进行数据拟合,获得了较为准确的拟合曲线,均方误差较小.
In order to solve the problem that simultaneous detection of temperature and strain in optical fiber sensing technology cannot be accurately obtained by using fiber multi-peak Brillouin scattering spectrum,the RBF neural network based on K -means clustering method is applied to the data fitting of optical fiber multi-peak Brillouin scattering spectrum. Firstly,the theoretical knowledge of Radial Basis Function(RBF)Neural Network and K -means clustering method is described. Secondly,it is concluded that different diffusion speeds will affect the fitting accuracy of the data by using RBF neural network algorithm to fit the data. However,the smoothness and accuracy of fitting curves cannot be guaranteed simultaneously. The RBF neural network based on K -means clustering method is used to fit the data and get more accurate fitting curve,and the mean square error is small.
作者
高原
高建军
杜佳豪
孔维宾
王如刚
周锋
Gao Yuan;Gao Jianjun;Du Jiahao;Kong Weibin;Wang Rugang;Zhou Feng(School of Information Technology,Yancheng Institute of Technology,Yancheng 224051,China;State Key Laboratory of Millimeter Waves,Southeast University,Nanjing 210096,China;Yancheng Optical Fiber Sensing and Application Engineering Technology Research Center,Yancheng Institute of Technology,Yancheng 224051,China)
出处
《南京师大学报（自然科学版）》
CAS
CSCD
北大核心
2019年第1期90-94,共5页
Journal of Nanjing Normal University(Natural Science Edition)
基金
国家自然科学基金(61673108)
东南大学毫米波国家重点实验室开放课题(K201731)
盐城工学院2018年优秀毕业设计(论文)培育项目
