The Research Of Multi-sensor Data Fusion Technology And Experimental In The Gas Detecting

With the rapid development of modern industry,air pollution is increasing quickly,which not only affects the air quality around,but also seriously affects the people's physical and mental health.Therefore,gas detection is the primary task to solve air pollution.Currently,gas sensor is most used to detect gas.But it is difficult to achieve with a single sensor to detect gas because of the cross-sensitivity of gas sensor.Therefore,the development of the multi-gas sensor system,which can monitor some gas composition and content of the environment in real-time,is particularly important.This paper reviews the current situation and development trend of research on multi-gas sensor array to detect the gas at home and abroad.After studying the working principle,characteristics and market price factors of various gas sensors,and the paper determines the research program of the multi-gas sensor information fusion technology.Finally,the detection system is built by four different gas sensors.Based on the system,we designed the structure of gas chamber and developed the circuits and the program for the multi-sensor signal detection.The hardware of system is based on the microcontroller of C8051F020,including the sensor conditioning circuit,power circuit,drive circuit and serial communication circuit.And the microcontroller program is developed in Keil.The software in PC is programmed by LabVIEW and MATLAB.The system has realized the functions of microcontroller and PC communication,data on the curve display and storage in LabVIEW.Artificial neural network is used to realize fusion algorithm of the collection of sensor information in the multi-sensor data fusion technology research.And the application is developed in the MATLAB platform.In the experimental part of the thesis,firstly,the stability and repeatability of the detection system are tested to provide a reliable guarantee for subsequent experiments.Then it analyzes the influence of gas flow experiments,and ultimately determines the best flow rate value of gas.Finally,the multiple information fusion algorithm is used on H₂S,NH₃ and styrene for qualitative and quantitative analysis.The experimental results show that the multi-gas sensor system can accurately identify the three types of gas.And the average error of quantitative analysis in NH₃,H₂S and styrene is 4.24%,4.02%,4.40%,indicating that the system basically meet the testing requirements of the subject.Finally,the paper summarizes the research work,and puts forward some suggestions for future research.