仿生嗅觉在空气净化器气味识别中的应用研究

Application research of biomimetic olfactory in odor recognition of air purifier

随着人们对高水平生活的追求,家用空气净化器的使用已越来越普遍。将仿生嗅觉应用于空气净化器领域,并探究其可行性。根据常见的家居气味(一氧化碳、硫化氢、醛类、苯类和TVOC等有害气味)以及一些常见的异味,选择合适的传感器,组成传感器阵列进行气味数据采集。实验选择了香烟、香水、榴莲、油漆作为气味源,对采集后的数据使用PCA算法进行特征提取,并建立气味指纹信息库,使用MATLAB软件建立BP神经网络进行仿真模拟。实验结果表明,该方法对四类气味都有着较高的识别率,可见此识别方法具有可行性。

With the pursuit of highlevel life,the use of household air purifiers has become more and more common.This study applies the biomimetic olfactory to the field of air purifiers and explores its feasibility.According to the common household smells such as carbon monoxide,hydrogen sulfide,aldehydes,benzene and TVOC and some common odors,selecting the appropriate sensors to form a sensor array for data acquisition.In this experiment,cigarettes,perfumes,durians and paints were selected as odor sources.The collected data were extracted by PCA algorithm,and the odor fingerprint information database was established.The BP neural network was built using MATLAB software to simulate.The experimental results show that the method has a high recognition rate for the four types of odors,which shows that the identification method is feasible.