Embedded neural network for fire classification using an array of gas sensors

Fire is one of the most common hazards in US households. In 2006 alone, 2705 people were killed due to fire in building structures. 74% of the deaths result from fires in homes with no smoke alarms or no working smoke alarms while surveys report that 96% of all homes have at least one smoke alarm. This study discusses the development of a fire sensing system that is not only capable of detecting fire in its early stage but also of classifying the fire based on the smell of the smoke in the environment. By using an array of sensors along with a neural network for sensor pattern recognition, an impressive result is obtained. Instead of confining the ANN to a PC based application, this work extends the implementation of the neural network fire classifier in a general purpose microcontroller. The result is a low cost intelligent embedded fire classifier that can be used in real life situations for fire hazards minimization, for example this intelligent fire classifier can be used for the development of a smart extinguisher that detects the fire, classifies it and then uses appropriate extinguishing material required for extinguishing the particular class of fire.