Data Driven Analysis Of Urban Traffic Situation And Digital Twins Modeling

With the increase of car ownership year by year,traffic congestion has become a common problem in major cities.Accurately judging the road traffic state and guiding drivers to avoid congested sections are important measures to alleviate traffic congestion and improve road traffic efficiency.The development of new generation technologies such as computer technology,artificial intelligence and big data provides new technical support and environment for the research and engineering practice of traffic state discrimination.This paper discriminates the road traffic state from the perspective of multi-source data,and considers the influence of different traffic data sources on the applicability of different traffic characteristic parameters.Firstly,this paper constructs the key data source selection system of different traffic characteristic parameters,and proves the rationality and accuracy of the system through example analysis,which effectively avoids the uncertainty of the solution result caused by the fusion of multiple data sources;Secondly,on the basis of the constructed index system,three evaluation indexes of average travel time ratio,average speed and average saturation are selected to construct a set pair analysis and evaluation model based on trapezoidal fuzzy number,which realizes the classification and accurate quantification of traffic situation.Through comparison with fuzzy comprehensive evaluation method and travel time ratio method,It is proved that the model proposed in this paper is fast,accurate and sensitive in situation evaluation;Thirdly,by using Python language and micro simulation software SUMO,the digital twin modeling of the studied road section is carried out,the automatic input of vehicles is realized,and the traffic situation discrimination model constructed in this paper is embedded in the simulation.And the road traffic situation is showing by dynamic curve.Through comparative analysis with fuzzy comprehensive model and travel time ratio model,the accuracy and sensitivity of the proposed algorithm in traffic situation discrimination are verified;Taking the flow of a certain period of a road section as an index,the simulation output results are compared with the real value.The results show that the average relative error of the simulation output is 8.79%,which shows that the data output of the digital twin road network is accurate and effective within the allowable range of error,which provides a certain solution for the repair of missing traffic data and the generalization of traffic data.