Design And Research Of Automobile Fire Alarm System Based On Multi-sensor Fusion

With the rapid development of China's economy,cars have become more and more popular in modern life,causing the car fire accidents occur frequently and irreparable damage.At present,the traditional way of detecting car fire is to use a single sensor to detect a single physical quantity,and the decision is very single.There are some problems such as high fault alert rate and high false acceptance,missing the best fire fighting time.Aiming at this problem,a multi-sensor fusion vehicle fire alarm system for vehicle developed,which possesses several functions such as automobile fire alarm,automatic breaking window and fire extinguishing,etc.By analyzing the mechanism and process of car fire,the vehicle fire detection features is determined,including temperature,smoke and carbon monoxide,and the overall design scheme of automobile fire alarm system was formulated.In the system,hardware,software and fusion algorithm are studied.Firstly,based on the modular design principle,the hardware part of the system is divided into seven independent modules,and then each module is selected and built in turn.Secondly,the design scheme of the system is proposed.The Code Warrior software is used to design the underlying computer program,and Lab VIEW is used to design the upper computer software.The system mainly realizes the functions such as data acquisition,feature quantity monitoring,data storage,linkage control and web publishing.Finally,the Bayesian fusion algorithm is applied to fire early warning and monitoring through the research on data fusion process,its principle and the common multi-sensor fusion technology.The data of fires in the simulated car compartment were collected and pre-processed,and the data fusion based on Bayesian network algorithm was carried out by using MATLAB software.The results show that the reliability of the three feature quantities fusion algorithm is higher compared with a single sensor.And then the The Netica software is used to construct the network model,and the collected data is verified and analyzed.On this basis,combined with the MATLAB fusion results,the discrete interval of the feature quantity of the car in different fire alarm states is determined.The rationalization of the proposed Bayesian network fusion algorithm is validated by experiments.