The study of multimodal gas recognition algorithm based on machine olfaction

摘要

Gas recognition, smell identification and source localization are among complex problems in today's industry. In this study, we employed an electronic nose (Enose) and applied the Locally Linear Embedding (LLE) algorithm to detect and classify four kinds of industrial gas including C02, NH3, CH4, Volatile Organic Compounds (VOCs). The AIRSENSE PEN3 Enose was used for gas detection and odor data acquisition. We compared the performance of the proposed LLE algorithm with the Principal Component Analysis (PCA) and Linear Discriminate Analysis (LDA) in extracting the gas characteristics as well as quantitative analysis and data processing. The results verified that the LLE algorithm outperformed other selected algorithms in multimodal gas recognition. Therefore, the LLE algorithm can play important role in the field of machine olfactory and odor identification.