Odor sensor using quartz-resonator array and neural-network pattern recognition

Abstract

An experiment of odor identification using a neural network is described. A quartz-resonator array with different coating films was used, and its output pattern was recognized using a neural network. Various odors of commercially available liquors were identifiable by the network following training. The sensing system was adaptive to environmental variations during the cyclic process of the data sampling and training. High recognition probability was maintained even under temperature variations.<>