Acquisition of fuzzy rules for fire judgment system

Recently, every building has fire alarm systems to detect a fire in its early stages and not to spread the damage of the fire. These systems are essential to protect human lives and properties. However, the lack of reliability in these systems, in which false alarms have arisen many times, has been a serious problem. This paper proposes a new intelligent fire judgment system with feature extraction from time series of smoke density using fuzzy rules acquired by Genetic Algorithm (GA). The GA in this paper uses selective elements method for rule generation. This system shows high reliability for the fire alarm systems through computer experiments. This paper also shows that effective features as fuzzy rules for each category scan be extracted using this method.