Study On Low-Altitude Atmospheric Monitoring Using Unmanned Aerial Vehicle System Equipped With Sensors

In recent years,China has implemented a series of effective measures to control air pollution,but the combined air pollution of fine particles and ozone is still severe.The current air monitoring measures in China have the problems of poor mobility and low ability to trace pollutants back to the source.With the maturity of UAV technology,the monitoring of atmospheric pollutants by the UAV platform equipped with sensors has attracted more and more attention.We conducted the studies on evaluation and calibration of sensor performance,construction and commissioning of the UAV pollutant monitoring system,development and testing of pollutant data transmission module from the UAV to the ground,and analyses of the vertical distribution characteristics of PM2.5,PM10,O₃in the atmosphere,in order to establish a reliable method for investigating the vertical distribution of atmospheric pollutants using the UAV pollutant monitoring system.We used the linear regression,the multivariate linear regression,the artificial neural network to model the data collected by the sensors for particulate matter and ozone,the reference instrument output data and the meteorological data（temperature,humidity and wind speed）.The comparison results of the goodness-of-fit,the root mean square error and the mean absolute error of the three fitting models showed that data of ozone and particulate matter by the sensors using ANN model were closer to the real values.The uncertainty analysis showed that the relative extended uncertainty of the O₃,PM2.5,PM10 by the ANN model met the instruction limit when the concentrations were 35 ppb,18μg/m³,28μg/m³.The results showed that the prediction accuracy of ANN model were 93%,90%and 89.7%for O₃,PM2.5 and PM10.On the basis of calibrating the sensors,we constructed the UAV pollutant monitoring system,including platform building,writing data monitoring codes for sensors and GPRS wireless communication module.Each module was integrated on the main control board,and was debugged on ground and on UAV at various vertical heights.The debugging results showed that the effect of airflow disturbance was lowest when the sensors were set up at 12 cm above the UAV rotor.The peanut shell software was used for the conversion of different network addresses in order to realize the external network access and the intelligent control of the monitoring system.The comparison of the UAV monitoring data with those from the State Controlled Mornitoring Site showed that the correlation coefficient between them was above 0.9.The UAV monitoring system was used to monitor the atmospheric PM2.5,PM10,O₃concentration from August to September 2020 at the height of 30 m,50 m,100 m in the atmosphere of Shenzhen University City Gymnasium,Xili Dasha River and Longhua East Industrial Zone.The results showed that at each height of the atmosphere,the daily average concentrations of PM2.5,PM10,O₃in University City Town,Xili,Longhua were 43-102μg/m³,5-43μg/m³,11-65μg/m³,respectively.With the increase in height,the concentrations of PM2.5 and PM10 tended to decrease,the concentrations of O₃tended to increase.The order from high to low concentrations of three pollutants at each height was as follows:Longhua>Xili>University City.The correlations of the concentrations of three pollutants between adjacent layers decreased with the increase of height.There was a significant negative correlation between PM2.5,PM10concentrations and ambient temperature as well as wind speed,and a significant positive correlation between PM2.5,PM10 concentrations and relative humidity.There was a significant positive correlation between O₃and ambient temperature.The correlations between O₃concentrations and relative humidity as well as wind speed were weak.