Olfactory classification using electronic nose system via artificial neural network

Abstract

Olfaction, according to modern research, has not yet been classified based on its known properties. Unlike the sense of taste and sight, the sense of smell does not have any known dimensions of category. Modern technologies of Electronic nose (E-nose) systems were used in analyzing smells. This study aimed to categorize different clusters of smell and differentiate their levels of reaction to an E-nose system comprising of different sensors. MQ - Metal-Oxide semiconductor gas sensors were used coupled with artificial neural network (ANN) using MATLAB to evaluate the systems capability of discrimination. Interfacing was done using Arduino Microcontroller for communication. MQ5 Gas sensor gave the most variance. This result confirmed that the system's ability to be used in future applications was suggested.