Application of fuzzy ART neural network to intelligent gas sensor arrays

Abstract

In this paper a new approach to gas identification using Fuzzy Adaptive Resonance Theory (ART) neural network and polymer coated surface acoustic wave (SAW) sensors array is reported. This method provides fast learning capability in the training mode and accurate gas classification in the sniffing mode. The method can be easily extended to gas mixtures as well as other sensor technologies. Since the Fuzzy ART neural network is self organizing classifier trained in unsupervised mode, it avoids the drawbacks associated with static feed forward neural networks.