Research On Deployment Of Mobile Pollution Source Telemetry Network On Urban Road Segments

In recent years,the number of motor vehicles in China has increased significantly,and their exhaust emissions have a great impact on the urban environment.In order to solve the various problems caused by exhaust emissions,the primary task is to measure various pollutants emitted by motor vehicles in the urban traffic network in a timely and accurate manner.The performance of the mobile pollution source remote sensing monitoring network,which can monitor the exhaust emissions of vehicles driving in the urban road network in real time,is positively correlated with the number of deployed remote sensing devices;the more the number of devices,the better the network performance.However,due to the expensive price and maintenance cost of remote sensing monitoring system,the application process of actual traffic network is greatly limited by the location and number of remote sensing devices,so there is an urgent need for scientific and effective location and deployment strategy to establish mobile pollution source remote sensing monitoring network.In this paper,we study the deployment of mobile pollution source remote sensing monitoring network,and investigate how to make reasonable location and deployment on urban road segments for different deployment purposes.For the deployment of mobile pollution source emission census on urban road networks,this paper proposed a remote sensing monitoring network deployment method based on the minimum cut.The method only needs known road network structure and traffic flow data,and firstly,the real urban traffic network is modeled.Secondly,the directed graph is searched by the depth-first search algorithm to obtain the directed loop road segments,the directed non-loop road segments,and the boundary deployment road segments.Finally,the location problem of remote sensing devices is transformed into solving the minimum cut problem,and the path segmentation algorithm based on the minimum cut finds the minimum cut set to deploy the devices on the directed loop road segments,and calculates the minimum road set for deploying the devices.The method considers all possible traffic routes and vehicles driving into or out of the deployed road network area,and obtains non-redundancy location results while ensuring that all on-road vehicles in the road network can be detected.For the deployment of spatio-temporal distribution prediction of mobile pollution source emissions on urban road segments,this paper proposed a high-order graph convolutional neural network for the deployment of mobile pollution source remote sensing network.Firstly,based on the environmental data and historical traffic data of very sparse remote sensing devices,a high-order graph convolutional network is used to predict the distribution of motor vehicle exhaust emissions at any moment on any road segment.Secondly,with the goal of improving the accuracy of motor vehicle emission prediction on urban road segments,an entropy minimization model based on greedy algorithm is used to mark the recommended priority of unmarked road segments locations by improving the accuracy of the prediction model according to the node incremental learning method.Finally,the road segments with the highest priority is used as the location of the new remote sensing monitoring node.The experimental results show that the proposed remote sensing monitoring network deployment method can effectively combine the spatio-temporal distribution prediction model of exhaust emissions with the remote sensing monitoring node deployment model,so that the new remote sensing monitoring nodes can maximize the accuracy of the spatio-temporal distribution prediction of motor vehicle emissions in urban traffic road segments.