Development And Research Of Intelligent Fire Detection System Based On Multi-sensor Information Fusion

In recent years,fire hazards to people's lives and property are generally on the rise.The highest proportion of casualties is caused by fires in residential buildings.At present,the scale of urban architectures is becoming larger and larger,the function is becoming more complex,and the intelligent level is continuously improved.In this context,how to eliminate the fire in the budding state has become the focus of urgent research.Currently,most of the fire detectors used in the market are based on a single sensor and combined with a simple algorithm to identify the fire.There is still high false alarm rate in the application.Therefore,this paper proposes an intelligent fire detection system based on multi-sensor information fusion,which aims to improve the timeliness and accuracy of fire identification.Aiming at the problem of high false alarm rate of a single sensor,a multi-sensor system is composed of smoke sensor MQ-2,CO gas sensor MQ-7,temperature sensor DS18B20 and infrared flame sensor.The lower computer module of the system is designed based on MCU,and the wireless communication between the upper computer and the lower computer is realized through CC2530 module.Based on Keil software and STM32 single-chip microcomputer as the core,the software system of the lower computer is designed,including the main program design,DS18B20 module program design,A/D acquisition module program design and communication module program design.Based on MATLAB and LabVIEW software,the upper computer is designed with PC as the core,realizing the functions of real-time display,data storage,data processing and real-time alarm.According to the theory of multi-sensor information fusion,the structure of the whole algorithm of fire detection system is designed.The single-chip microcomputer is used to pre-process the collected sensor signal,such as limiting and normalizing,to complete the information layer fusion of the system.Then the pre-processed signal is sent to the upper computer.Based on the MATLAB software,the design of BP neural network fusion device and fuzzy inference system is completed,and the feature layer fusion of the system is completed through the processing of the upper computer to the bottom data.According to the result of feature layer fusion and the duration of fire,a multi-level criterion is generated to get the results of fire detection in the environment and the fire alarm is realized.So that the decision level fusion of the fire detection system is complete.At present,the research of intelligent fire detection technology is mostly accomplished by simulation.Finally,the collection of fire samples,fire response time experiment and anti-interference experiment are carried out.The fire samples provide the basis for the training of neural network and the design of fuzzy inference system.Through the fire response time experiment and comparing with an independent smoke-temperature composite fire detector on the market,the timeliness of fire detection system's response to fire is illustrated.Through the anti-interference experiment,the good anti-jamming ability of the fire detector system is verified.