Study On Home Fire Monitoring System Based On Fuzzy Neural Network

In the field of security,fire detection technology has always been an important research topic.The traditional method of fire detection is relatively simple.It is based on a simple mathematical model and a single sensor.The degree of intelligence is weak.It only detects single physical information(such as smoke or temperature),while the kitchen smoke,air fog,smoking,etc.Alarms will appear,there are false positives and other false positives.Therefore,given the shortcomings of the previous fire detection technology and the limited application in the field of safety,it is of great value and practical significance to develop a fire detection device with high intelligence,automatic alarm,and strong identification capability.Improve and improve all aspects.For this purpose,the research in this paper mainly includes the following work:1.Using a multi-source information fusion technology,a fuzzy monitoring algorithm and a Levenberg-Marquardt BP neural network algorithm are used to fuse,calculate,and judge temperature,smoke,and CO physical quantities at the time of fire,thereby constructing a fire monitoring system.The traditional BP algorithm is compared with the LMBP algorithm,so that fuzzy control and LMBP neural network are selected as the data fusion algorithm,and the results are subjected to multiple simulation tests.The traditional BP algorithm model is improved,the running speed of the algorithm is improved and the executability in the computer is achieved,and the decision of whether or not the fire occurs is finally obtained.2.Using multi-sensors to collect multi-dimensional information,monitoring and analysis of the three physical quantities of smoke,temperature,and CO gas at the time of fire.The traditional fire detection method(ie,the collection and analysis of singlephysical quantity)has been improved,making it more comprehensive to monitor fire signals.3.The fire monitoring system uses ARM series CortexA8 embedded processor,select,install,configure temperature,smoke temperature and gas sensor module,use C and C++ language,transplant Linux operating system,write device drivers.It improves the features of traditional single-chip microcomputers(C51,AVR)that are single-function and slow in data processing,especially for the complexity of fire data,making the data more efficient and accurate.Finally,the fuzzy neural network algorithm was transplanted into the hardware.Through the ARM hardware platform test,the result of accurate fire prediction was obtained.The article completed simulation and physical testing of intelligent fire monitoring.The experimental results show that using LMBP neural network algorithm can improve the processing speed of fire data more than BP algorithm,and using multi-sensor technology to collect information improves the accuracy of fire monitoring.The research of the intelligent fire protection system has provided security for people's lives,improved the instability of traditional fire protection systems,reduced the interference caused by external environmental factors,and has important research significance and application prospects.