Design And Implementation Of Community Fire Monitoring System Based On Wireless Sensor Network

With the increasing number of residential communities and the improvement of people’s safety awareness,more and more attention has been paid to the safety of fire fighting system.In recent years,the demand of fire detectors is increasing year by year.In view of this situation,this thesis designs a fire monitoring system suitable for community buildings to meet the needs of fire monitoring and alarm.Through the overall analysis of the functional requirements of the system,the ZigBee networking technology is selected to build the system after comparing various communication modes.After analyzing protocol system,characteristics and topology of ZigBee technology,the tree network is designed,and the overall scheme and implementation process of the fire monitoring system are designed.Firstly,the software and hardware of three kinds of ZigBee function nodes,STM32-SIM900A gateway and fire monitoring center are designed respectively.In terms of hardware,the power amplifier circuit,sensor circuit and power circuit of terminal node are designed.And the interface circuits of coordinator,SIM900A module and gateway controller are designed.They lay the hardware foundation for the fire monitoring system;In terms of software,the programs of three kinds of ZigBee function nodes and sensors are designed to upload the environmental data to the gateway.The software design of STM32-SIM900A gateway ensures that the data are transmitted between different protocols and finally uploaded to the upper computer.The fire monitoring center mainly involves the design of SQLite database and the program design of user interface to achieve login,real-time data display,node configuration,historical data query,fire alarm and other functions.Secondly,based on the selection of fire environment parameters and the analysis of fire detection algorithm,in order to make the fire alarm more reliable and accurate,the global fish swarm algorithm is used to optimize the initialization weight and threshold value of BP neural network.Then the BP neural network algorithm is used for fire prediction,including model building,parameter setting and optimization process.And it is compared with individual BP neural network algorithm for simulation.The results show that the optimized algorithm has more advantages in iterations,convergence speed and error.Finally,the fire monitoring system is tested and analyzed.The lower computer includes serial port test of sensor data collection and ZigBee node networking.The STM32-SIM900A gateway includes network assistant test of wireless sensor network data.The upper computer includes real-time data display,historical data curve display,short message alarm function test,etc.Due to the obstacles in the community building such as walls,the interference of WiFi and packet loss rate are analyzed.