The Adaptive Modeling Method Research Of Urban Regional PM2.5 Concentration

Nowadays,many cities are under severe fine particulate matter PM2.5(Particulate matter with diameters less than 2.5 um)pollution,high PM2.5 concentration is endangering human health,and damaging the ecological environment.So fine-grained urban PM2.5 concentration distribution information is of great importance to protect people's health and control air pollution.Urban PM2.5 concentration distribution has big regional differences and obviously affected by environment.Currently,urban PM2.5 concentration is monitored by a few fixed monitoring sites,this method cannot provide fine-grained PM2.5 concentration distribution,there is a big difference between the actual pollution of the people's surrounding environment and official provided data.PM2.5 varies in urban spaces non-linearly and influenced by multiple factors,such as meteorology,traffic volume and geographical feature.In this paper,we infer the real-time and fine-grained PM2.5 information,based on the original PM2.5 data and a variety of data in the city,such as POIs data,meteorology data,traffic flow data and geographical feature data.To ensure the data coverage and get the fine-grained PM2.5 concentration distribution,the original PM2.5 data in this paper is collected by sensors built-in mobile taxis.This paper collected the city datasets more than a year,and analyzed the relation between the PM2.5 data and regional function,traffic,weather conditions.Statistical analysis and correlation analysis result show that the distribution of PM2.5 concentration in the area is complex,and affected by related factors.It has proved that an official monitoring site in the testing area cannot provide accurate PM2.5 concentration distribution.This paper divided the testing area into grids and assumed the PM2.5 concentration in a grid w uniform(while different grids may have different results),the collected PM2.5 data was merged into each grids according to its positional information.To improve the flexibility and efficiency of the system,on the premise of guarantee system's accuracy,this paper proposed a grid refinement criteria,and dynamically adjust the grids resolution.For undetected grids,this paper simulated the spread of fine particulate matter in the form of probability transfer based on the idea of random walk.An adaptive resolution-probabilistic concentration estimated method(AG-PCEM)is proposed to infer PM2.5 concentration with adaptive resolution in testing area.This paper selected the easy deployment PM2.5 sensor and the precise dust particle detector as a calibration device,through the experimental analysis,considering the accuracy of the prediction results and the system efficiency,this paper selected the optimal grid resolution and adopted 1 hour monitoring frequency and 5 monitoring vehicles as monitoring strategy.This paper gradually collected the measurements throughout a year.Experimental data has verified that the proposed method can achieve good performance in terms of computational cost and accuracy.The computational cost of AG-PCEM is reduced by about 40.2%compared with a static grid method PCEM under the condition of reaching the close accuracy,and the accuracy of AG-PCEM is far superior as widely used artificial neural network(ANN)and Gaussian process(GP),enhanced by 38.8%and 14.6%,respectively.The system can be expanded to wide-range air quality monitor by adjusting the initial grid resolution,and our findings can tell citizens actual air quality and help official management find pollution sources.