Early Detection of Fire Hazard Using Fuzzy Inference System

Fast detection of fire allows to protect human's health and life as well as avoid loss of property. Most fire sensors have simple construction, consisting of one or two sensing elements (usually detecting smoke and temperature) and an uncomplicated algorithm. This paper presents an idea of using fuzzy inference for early fire detection based on simultaneous analysis of data from multiple sensors. The measurement system includes a head with two accurate sensors: electrochemical carbon monoxide sensor and resistance temperature sensor. The implemented fuzzy inference algorithm accepts as inputs the values measured by the sensors, signal rise rate and signal variability. The main goal of the system is to early detect fire hazard or sensors malfunction. A series of experiments and detailed analysis of results were performed. It has been proved that fuzzy inference is suitable to the presented application.