A review of algorithms for SAW sensors e-nose based volatile compound identification

Olivier Hotel (1) , Jean-Philippe Poli (1) , Christine Mer-Calfati (2) , Emmanuel Scorsone (2) , Samuel Saada (2)

1 LI3A (CEA, LIST) - Intelligence Artificielle et Apprentissage Automatique (CEA, LIST)
2 LCD (CEA, LIST) - Laboratoire Capteurs Diamant (CEA, LIST)

Recent advances in odour sensors have led to the development of new applications; among them, electronic noses have gained major interest and found successful applications in many fields. An electronic nose is a device composed of an array of odour sensors with sensitivity to a wide range of chemical compounds. Reliable electronic nose systems rely on advanced data processing techniques. Among them, machine learning has become a core technique for electronic nose design. In this document, we describe several machine learning algorithms and compare their performances on different features used in state of the art electronic nose systems.

Mots clés: instrumentation; diamond; signal processing; fuzzy logic; online learning; artificial intelligence; Machine learning; Data processing; SAW sensors; Electronic nose; Odour recognition; trustworthy artificial intelligence; INDEPENDENT COMPONENT ANALYSIS; SUPPORT VECTOR MACHINES; PATTERN RECOGNITION; NEURAL NETWORK; CLASSIFICATION

Domaines: Chimie analytique; Machine Learning [stat.ML]; Micro et nanotechnologies/Microélectronique

Olivier Hotel, Jean-Philippe Poli, Christine Mer-Calfati, Emmanuel Scorsone, Samuel Saada. A review of algorithms for SAW sensors e-nose based volatile compound identification. Sensors and Actuators B: Chemical, 2018, 255, Part 3, pp.2472 - 2482. ⟨10.1016/j.snb.2017.09.040⟩. ⟨hal-01879582⟩