P2.9.14 An artificial immune system model for gas sensors drift mitigation

Event

14th International Meeting on Chemical Sensors - IMCS 2012
2012-05-20 - 2012-05-23
Nürnberg/Nuremberg, Germany

Chapter

P2.9 Technology and Application

Author(s)

G. Magna, E. Martinelli, A. Catini, A. D'Amico, C. Di Natale - Department of Electronic Engineering, University of Rome "Tor Vergata" (Italy), S. De Vito, G. Di Franci - Italian National Agency for New Technologies, Energy and Sustainable Development (ENEA) (Italy), A. Vergara - BioCircuits Institute, University of California San Diego (USA)

Pages

1737 - 1740

DOI

10.5162/IMCS2012/P2.9.14

ISBN

978-3-9813484-2-2

Price

free

Abstract

Nature is a continuous source of inspiration for problem solutions. To this regard, paradigms of the immune system have recently been implemented in several applications, where they show interesting potentialities and promising results. Herewith, an adaptive classification model inspired by algorithms modeled on the human immune system is introduced and applied to an array of gas sensors. The algorithm is found to be efficient to preserve gas recognition capabilities even in presence of significant drift. This model offers a compact and adaptable configuration to the time evolution of the sensors signals distribution with respect to the traditional classification models. The performance of the algorithm has been evaluated with artificial and experimental datasets, and the classification rates have been compared with those exhibited, on the same datasets, by standard classifiers.