Smart Electronic Nose: An Empirical Evaluation of New Multi-channel Gas Sensors for Smell Classification

Nowadays there is a great interest in the development of devices capable of differentiating between different smells either for quality control in food or beverages or for the detection of smells in enclosed spaces and to be able to act accordingly. This paper presents a case study for the evaluation of new gas sensors for aroma identification. A prototype has been developed using gas sensors based on carbon nanotubes and other commercially available multichannel sensors. A case study has been proposed for the classification of aromas of different beverages and foods with the aim of obtaining a first dataset to evaluate its performance. From the dataset obtained, different pipelines have been applied including data augmentation processes, dimensionality reduction and different classifiers have been applied. Finally, the results obtained are presented and discussed. The dataset collected in this work can be used for future research in this area with the sensors used.