厚膜SnO_2气体传感器的嗅觉特征提取与处理  

Processing and picking - up the olfaction character of thick film SnO_2 gas sensors

作　　者：邵应清[1] 邹小波[2] 

机构地区：[1]江苏理工大学汽车学院,江苏镇江2120132 [2]江苏理工大学生物与环境工程学院,江苏镇江212013

出　　处：《传感器技术》2000年第6期15-17,共3页Journal of Transducer Technology

摘　　要：用一组厚膜ＳｎＯ２气体传感器阵列模拟人的嗅觉形成过程，对５种不同体积分数乙醇溶液进行分 析。详细叙述了实验过程，分别从每个气体传感器与气体反应的曲线中提取４个特征，用ＢＰ神经网络对 样本特征值的处理，对不同体积分数乙醇溶液进行识别。神经网络对训练集的回判正确率为１００％，对测 试集测试正确率为９０％。An thick - film SnO2 gas sew array used to imitatethe function of human olfaction system is used to analyze five different concentrations of ethanol solution. A detailed exposition is given to the process of the experiment. Four characters from each curve of the gas sensor's reaction to gas are picked - up. BP(Back - Propagation) neural network recognition are used to identify the samples from the five different ethanol solutions. The recognition probability of the network is 100% to the training set and 90% to the testing set.

关 键 词：气体传感器 嗅觉特征提取 厚膜 二氧化锡 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]