Identify household burning smell using an electronic nose system with artificial neural networks

This paper presents the ability of a new electronic nose (EN) system to identify various sources of burning smell. The EN has been developed based on the concept of human olfactory system by using various kinds of metal oxide gas sensors (MOGSs) as the olfactory receptors. The headspace of the EN is put directly over the tested smell and the time series signals during the MOGSs absorbing the smell are collected. By controlling the temperature and the humidity inside the tested chamber, the signals data from the same source of burning smell in every repetition data are highly correlated and each source of burning smell has a unique pattern of time series data. Therefore, the error back propagation neural network (BPNN) is able to identify 99.6% of the tested data accurately by using only a single training data from each source of smell. The results show the high possibility to apply the EN as a reliable fire detecting system.