Independent Component Analysis and Neural Network Applied on Electronic Nose System

Electronic noses are being developed as systems for the automated detection and classification of odors, vapors, and gases. Based on the study of the theory and constitutes of the electronic nose system, a set of independent component analysis (ICA) algorithms with BP neural network, for detection of gas mixture is designed and constructed, and the data processing which is measured by an electronic nose system consisting of five gas sensors is carried out. The results show that ICA algorithm can make a good classification for the data and reduce the data correlation. As the input of the BP network, it can predigest the structure and improve the convergence speed of the network.