Temporal And Spatial Simulation Research On Road Traffic Exposure Of Urban Rainstorm And Waterlogging Based On Trajectory Data

Affected by factors such as global climate change and urbanization,rainstorms and waterlogging have become a common natural disaster in cities,posing a serious threat to the sustainable development of cities and people’s lives.The urban road traffic network is an important guarantee for people’s travel activities and social development.Once it is hit by heavy rain and waterlogging disaster,the traffic will be abnormal and complicated,which will have a very negative impact on the normal operation of the urban system.So the question is that how to quantify the exposure of road traffic in rainstorm and waterlogging and dynamically simulate its evolution law in rainstorm events is of great significance for reducing road traffic risks.Therefore,this research deeply analyzes the impact characteristics of road traffic exposure and time-space changes under heavy rain and waterlogging disasters,constructs a road traffic exposure calculation model,and proposes a dynamic simulation method of road traffic exposure based on multi-agents.The main research contents and results are as follows:(1)Summarize the basic concepts about road traffic exposure,and analyze the impact of disaster-causing factors,disaster-pregnant environment and disaster-bearing bodies on road traffic exposure in urban rainstorm and waterlogging disasters.Using data mining methods,the correlation between road traffic exposure characteristics and traffic abnormalities is analyzed.Based on the results of data analysis,a formal expression model of road traffic exposure under the rainstorm and waterlogging scenario is proposed.(2)Based on the feature expression model,relevant features are extracted and processed from the trajectory,hydrological and basic geographic information data sets,and a quantitative calculation structure including direct and indirect road exposure is constructed.The random forest algorithm was used to identify the road with water,and the direct exposure of road traffic is quantified.Using the method of feature engineering in machine learning for reference,the index weights such as road grade,length and DEM are determined,and the quantitative calculation of indirect exposure of road traffic is completed.(3)Analyze the behavior characteristics of driving vehicles in rainstorm and waterlogging disasters,propose a dynamic simulation method of road traffic exposure based on a multi-agent based model,design the related rules and interaction methods of the agent,and create different scenarios like rainstorm and waterlogging scenarios and intervention scenario,and then analyze the different results.The experimental results show that the dynamic simulation method designed in this paper can accurately describe the temporal and spatial changes of urban road traffic in rainstorm waterlogging,which is conducive to the emergency department to take targeted measures to reduce the urban traffic risk.