Design of an electronic nose with artificial neural networks to determine fruit ripening

Abstract

Electronic noses, due to their portability, robustness and low price, are currently extremely useful in the agri-food area, allowing products to be analyzed in a non-destructive way, monitoring results in real time. In this research, a study is carried out based on a design of an electronic nose to naturally accelerate the ripening of climacteric fruits using ethylene gas produced by them. Artificial neural networks were used with a sequential model and binary classification with the ADAM optimizer. The results show a 100% precision to determine the ripeness of the avocado when it receives the gases produced by other fruits such as apples and bananas.