Design And Research On Fire Detection And Alarm System Based On Multi-sensor Information Fusion Technology

With the rapid development of national economy and the accelerating ofurbanization level, high-level residential buildings and large-scale public buildings aregetting more and more, followed by fires and losses.For preventing the seriousconsequences of fires, the timeliness and accuracy of fire detection is getting more andmore important. On the basis of an in-depth analysis on the research status anddevelopment trend of domestic and foreign fire detection and alarm system, the papermainly makes the research on the fire detection and alarm system based on thecomposite detection and intelligent algorithm.First of all, a distributed fire detection alarm system, which is made up ofcomposite detection node and fire alarm controller, is designed to meet the functionalrequirements. For micro-controllers of the major nodal points, chips of STM32F103areadopted. In order to solve these problems, such as controllers overburdening, slowresponse and low reliability, the composite detection node not only collects three fireparameters like the temperature, CO concentration and smoke concentration, butachieves the information processing and judgment. Fire alarm controllers make it cometrue that each node can be inspected and information can be displayed and stored.Secondly, for the traditional fire detection technology is based on the single fireparameters and it causes false and failure alarm, so the paper focuses on the intelligentfire detection system on the basis of fuzzy theory and neural networks. Using thetemperature, CO concentration and smoke concentration as the inputs, the systemcarries out the information fusion of mutil-sensors by means of intelligence artificialalgorithm. The system adopts the method of fuzzy logic and neural networks. It firstly uses the fuzzy logic to make the fuzzification, inference machine and defuzzification ofthe inputs of the three systems and sets up64fuzzy rules. Then it adopts neuralnetworks models to make the internet practice. Finally, it carries on the organintegration, makes the judgment by means of fuzzy logic at the policymaking level andachieves algorithm emulation through MATLAB.This paper aims to design and realize fire detection and alarm system, which isbased on multi-sensor information fusion technology, and accomplish the design andrealization of the hardware of the fire composite detection node and fire alarm detector.What’s more, it even uses the fusion algorithm of multi-sensors data and conducts thesimulations so that the system can makes timely and accurate judgment about whetherthere is fire. Therefore, it has the strong ability to make the fire prediction and alarm.