Odor sensing system using neural network pattern recognition

Author

Moriizumi, Toyosaka ; Nakamoto, Takamichi

Author_Institution

Dept. of Electr. & Electron. Eng., Tokyo Inst. of Technol., Japan

fYear

1992

fDate

9-13 Nov 1992

Firstpage

1645

Abstract

The development of an odor sensing system is described. The system consists of plural gas sensors, which are quartz resonators coated with various kinds of adsorption films, and a computer for the pattern recognition of the sensor outputs (resonance frequency changes due to odorous molecule adsorption). The error backpropagation algorithm is used for the pattern recognition. Aromas of whisky, fragrance, and flavor were successfully indentified, and the flavor difference due to a small amount of a foreign substance was also distinguished, demonstrating a detection limit approximating that of normal persons

Keywords

backpropagation; chemical variables measurement; computerised instrumentation; gas sensors; neural nets; pattern recognition; adsorption films; chemical variables measurement; computerised instrumentation; development; error backpropagation algorithm; flavor difference; gas sensors; neural network; odor sensing system; odorous molecule adsorption; pattern recognition; quartz resonators; resonance frequency; Artificial neural networks; Gas detectors; Neural networks; Olfactory; Pattern recognition; Resonance; Resonant frequency; Sensor arrays; Sensor systems; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Industrial Electronics, Control, Instrumentation, and Automation, 1992. Power Electronics and Motion Control., Proceedings of the 1992 International Conference on

Conference_Location

San Diego, CA

Print_ISBN

0-7803-0582-5

Type

conf

DOI

10.1109/IECON.1992.254352

Filename

254352