基于岭回归的气体预警穿戴系统的灵敏度校正  

Sensitivity correction in wearable gases alarm system based on ridge regression

作　　者：袁方红 许武军[1,2] 赵海森 

机构地区：[1]东华大学信息科学与技术学院,上海201620 [2]数字化纺织服装技术教育部工程研究中心,上海201620

出　　处：《微型机与应用》2017年第17期79-81,85,共4页Microcomputer & Its Applications

摘　　要：在气体预警穿戴系统中,使用电化学气体传感器来采集环境中CO气体的浓度。由于电化学气体传感器的灵敏度随着温度有着非常显著的变化,为了提高传感器的检测精度,采用一种基于岭回归算法对电化学气体传感器进行灵敏度校正。开发算法的软件工具采用Python语言进行编程,根据传感器和温度关系的已有关系样本对其进行拟合,并对拟合结果进行交叉性验证分析。通过分析对比可知,基于岭回归算法的电化学气体传感器灵敏度曲线拟合效果好于普通的最小二乘法拟合。The electrochemical gas sensor is used to get the concentration of CO in wearable gases alarm system. Science the sensitivity of electrochemical gas sensor changes much more when the environment temperature has changed,and it will have a bad effect on the accuracy of electrochemical gas sensor. So this paper presents a method which is based on ridge regression to regulate the sensitivity of electrochemical gas sensor. The software which we use is python. And we carry out linear fitting according to the relationship between the sensor and the temperature from the samples,and the fitting results are cross checked and analyzed. The results show that the sensitivity curve of the electrochemical gas sensor based on the ridge regression algorithm is better than that of the ordinary least squares fitting.

关 键 词：气体预警穿戴系统 电化学气体传感器 灵敏度校正 岭回归 PYTHON 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]