Classification of Food Types in a Box with Gas Sensors Using a Machine Learning Method. Case Study of Intelligent Electronic Nose

Abstract

The Active and Assisted Living (AAL) paradigm has helped the expansion of the use of sensors, which is increasingly common in research work related to activity monitoring. However, one of the major disadvantages found in using sensors inside the home or other types of environment to monitor activities is the non-acceptance by users of having sensors around them because they may see their privacy compromised. Therefore, nowadays it is important to search for non-invasive and low-cost sensors that provide the user with the security and accessibility to feel comfortable with them. In this work a case study has been carried out with the design and construction of a Metal Oxide Semiconductors (MOS) sensor array with which it is intended to monitor the type of food used in the kitchen by means of the K-Nearest Neighbours (K-NN) machine learning method. Specifically, the case study presented seeks to differentiate between bananas, lemons, chorizo and prawns in a first approach as an intelligent electronic nose. Gas sensors have been used to take advantage of their non-invasive character for the user, although the disadvantage of not being widely explored in the scientific literature has been found. The results obtained in the case study presented in this paper to classify these four foods have been promising to advance in this research topic.