摘要
针对传统的A/D值转换物理量回归方法中存在表达不统一、动态适应性弱和在线非线性校正能力不足等问题,尝试将机器学习的ELM网络引入到该应用中。在分析A/D值转换物理量回归的知识要素基础上,依托ELM网络的非线性映射能力,提出利用遗传算法优化ELM网络,并利用其实现统一数学表达的A/D值转换物理量回归方法。实际应用表明,该方法对物理量回归问题可实现统一的数学模型表达,泛化性好,且非线性校正能力强,实现了各类A/D值转换物理量回归应用。
In view of the problems of traditional A/D value conversion physical quantity regression method,such as inconsistent expressions,poor dynamic adaptability and insufficient online nonlinear correction ability,the ELM(extreme learning machine)network of machine learning is introduced into this application. On the basis of the analysis on knowledge elements of A/D value conversion physical quantity regression,and relying on the nonlinear mapping ability of ELM network,an A/D value conversion physical quantity regression method is proposed,which uses genetic algorithm to optimize ELM network,so as to achieve consistent mathematical expressions. The practical application shows that the method can be used to achieve consistent mathematical model expressions for physical quantity regression and are of good generalization. In addition,it is of excellent nonlinear correction ability,so the regression application of various A/D value conversion physical quantities can be realized satisfactorily.
作者
王平
王宜怀
刘长勇
彭涛
WANG Ping;WANG Yihuai;LIU Changyong;PENG Tao(School of Mathematics and Computer Science,Wuyi University,Wuyishan 354300,China;School of Computer Science and Technology,Soochow University,Suzhou 215006,China;The Key Laboratory of Cognitive Computing and Intelligent Information Processing of Fujian Education Institutions,Wuyishan 354300,China;Collaborative Innovation Center of Novel Software Technology and Industrialization,Suzhou 215006,China)
出处
《现代电子技术》
北大核心
2020年第17期141-146,共6页
Modern Electronics Technique
基金
国家自然科学基金资助项目(61672369)
中央引导地方科技发展专项资金项目(2018L3013)
福建省教育厅科研基金资助项目(JA15522)
武夷学院校科研基金资助项目(xl201016)
福建省本科高校教育教学改革项目(FBJG20190281)。
关键词
机器学习算法
模/数转换
极限学习机网络
遗传算法
优化方法
物理量回归
动态校正
machine learning algorithm
A/D conversion
ELM network
genetic algorithm
optimization method
physical quantity regression
dynamic correction
