Nir Katchinskiy, Greeshma Kayala, Katherine Liu, Daniela Villasmil, and Mrinal K. Mandal
Electronic Nose, Feature Extraction, Gas Sensors
The Electronic Nose is a device used to distinguish between scents. The hardware portion of the electronic nose is comprised of an array of eight different gas sensors installed in an open chamber. Scents were characterized using different features extracted from sensor response curves, and then recognized by finding the minimum Euclidian distance in multi-space between the features. The final product is generic in its applications, compact, portable, adaptable to future development and user friendly. The resulting product has an operation of 2.5 minutes per sample and the total clearing time between samples is 1.75 minutes on average. The results show that the data recognition ability of the E-nose improved as the number of samples of a given scent increased. A recognition success rate of 95% was achieved.
Important Links:
Go Back
