Wearable electronic nose for human skin odor identification: A preliminary study.

Highlights • An approach for detecting human skin odors via wearable electronic nose is studied. • A sensor array based on CNTs/organics composites is applied for diverse VOCs sensing. • Contrast response is extracted and computed by PCA and K-means to identify different odors. Abstract Human skin odor is an important pheromone that can serve as an indicator for physiological status, or as a "fingerprint" for individual identification. It is a complex task to identify the organic components that constitute skin odor. An electronic nose is proposed to identify skin odors. This paper consists four parts. First is to prepare a gas sensor array, two functionalized carbon nanotubes (CNTs) are encapsulated by three different polymers to form six sensing materials, and they are deposited on a substrate to form a sensing array. A circuit is designed for data acquisition and wireless transmission. Second, gas chromatograph-mass spectrometry (GS-MC) is carried out to demonstrate the organic components of skin odors. Then, humidity and gas sensing properties of the sensor array are evaluated and tested in the laboratory. Third, human study is designed and conducted for monitoring of skin odor using the prepared electronic nose. The last is signal processing, to achieve highly accurate odor identification, contrast response is extracted as the feature. The best clustering accuracy, which is computed by unsupervised K-means, is 91.67%. This paper represents an important step in the research and development of electronic nose for human skin odor identification. [ABSTRACT FROM AUTHOR]