Research On Smart Home Fire Detection System Based On Fuuzy Neural Network

With the rapid development of science and technology and people livingstandard improved, more and more electrical equipments were used in the daily life,and they are more and more advanced. The concept of Smart Home also comes intopeople’s life. In the Smart Home environment, fire detection is an essential work. Butmany research of fire detection at home and abroad have focused on large sites atpresent. So the development of a kind of economical and practical, efficient andaccurate smart home fire detection system has great significance.Smart home fire alarm system is aimed at detecting fire, the most important keyis real-time and accuracy, the loss of life and property will be less. In present, most ofthe fire detection systems are based in a single smoke sensor method, although thismethod can detect the occurrence of fire, but there is a high rate of false positives,such as the kitchen smoke and smoking can cause wrong alarm. In this paper, threesensors such as temperature sensor and smoke sensor and CO gas sensor are used inorder to improve the accuracy and real-time of detection. Then using fuzzy neuralnetwork algorithm to collect information and classification of the three sensors,finally got the result of the fire occurred or not. In this paper, we achieved initialprogress in study the intelligent household fire detection system, which based onfuzzy neural network. The host controller is a single chip microcomputer,CC2530F256is used as wireless data transmission module, data is analyzed with itsinternal8051microprocessor, the text message is sent to a specific phone numberthrough the GSM network.The paper makes a deep analysis and research on related theory and technology，including fire detection technology and multi-sensor information fusion technologyand fuzzy system and neural network and the ZigBee wireless communicationtechnology. Hardware of the smart home fire alarm system is designed, mainlyincludes the main controller and the fire detectors. And the algorithm of the smarthome fire detection system is designed, using fuzzy neural network fusion algorithm as the system modeling algorithm. Through the learning and training of the firesample data, simulation tests is carried out at a specific home environment and thesimulation results is summarized and evaluated, further refined and improved of thisstudy is proposed.