Machine learning-integrated biomimetic electronic noses: Future perspectives

Abstract

The electronic nose (E-nose) is an innovative device that mimics the human sense of smell. E-noses are used for effective detection and discrimination between complex odors. Compared to traditional odor detection methods, E-nose technology employs a sensor array that differentiates and measures airborne smells through a combination of electrical signals generated by the sensor array when detecting odors. In addition, the incorporation of machine learning for data processing has enhanced the sensitivity and selectivity of odor molecular detection. However, certain limitations exist, such as a limited range of detectable odor molecules and low analytical accuracy for similar compounds, which challenge the claim that E-noses can fully mimic human olfaction. In this paper, we provides a general overview of the E-nose structure and its operating principles, as well as a summary of recent research and practical constraints in detecting volatile organic compounds. Moreover, this review paper discusses the future development of biomimetic E-noses in conjunction with other technologies and describes their potential commercial applications, including through E-commerce platforms. The critical review contributes to the E-nose literature by offering insights into how the E-nose device can solve real-world problems and by proposing directions for future advancement.