Machine learning-assisted electronic nose and gas sensors

In this chapter, an electronic nose (E-nose) will be presented as the first application of machine learning (ML). Several gas sensors have been identified and developed in the past, showing limited specificity to various gaseous chemicals. However, when used collectively as a set, gas sensors can be utilized to identify and quantify specific gaseous chemicals, especially with the help of machine learning (ML) enabled data analysis. ML can analyze the high dimensional gas sensor data and classify the type of gaseous chemical and its concentration. The process mimics the mechanism of the human/animal nose, where the olfactory receptors correspond to the gas sensor array and the ML algorithm to the data processing in the brain. Hence, A ML-enabled gas sensor array is called an electronic nose or E-nose. Various types of gas sensors will be explained first, followed by the ML techniques used in recent literature to make E-nose functional.