A Portable E-Nose Endowed With Subjective Evaluation Function of Air Quality in Vehicles

Abstract

In-vehicle air pollution, especially the volatile organic compounds (VOCs) pollutants, is an important source of passenger discomforts and consumer complaints. In this work, we designed a portable instrument based on an electronic nose (e-nose) to address the task of subjective assessment of in-vehicle air quality (IAQ). Our instrument consists of three modules: 1) the preconcentrating module that concentrates the VOCs to detectable level through adsorption and desorption; 2) the sensing module that measures the odor fingerprints of the target gas with an array of metal oxide semiconductor sensors; 3) the processing module that classifies the measured responses into the corresponding subjective grades via a stacking-based ensemble learning method. A dataset for subjective IAQ assessment was built to train and test the e-nose, which contains gas sample responses measured by the instrument and labels obtained by jury tests. Grading experiment results validated the practical performance of the proposed instrument as a direct case of narrowing gaps between objective assessments and subjective perceptions based on instrumentation.