Research On Traffic Flow Characteristics Under Severe Weather Conditions

Digital transportation integrates advanced information technology and intelligent transportation,and its development is inseparable from the drive of traffic flow data.Indepth analysis of the characteristics of traffic flow data plays an important guiding role in traffic planning,traffic design,traffic control and other aspects.In the operation of road traffic,the impact of road traffic environment and traffic accidents on traffic flow is particularly prominent.Full consideration of the impact of environmental factors in traffic flow prediction research can provide important theoretical support for the realization of in-depth research and application of digital traffic technology.The paper uses the traffic flow data provided by the British Highway Administration and the meteorological factor data obtained from the Weather underground website as the core research data.First of all,it analyzes the time changes of traffic flow and the composition of vehicles on the British M3 highway and A3 ordinary roads,and conducts an in-depth analysis of the average vehicle speed under different rainfall intensities and the changes in the flow of different models,using the classic Van-Aerde traffic flow The model calibrates free flow speed,capacity,blocking density and critical speed,and obtains the changes of traffic flow parameters under rainy weather and normal weather.Secondly,considering the influence factors of traffic flow such as temperature,humidity and wind speed,the correlation analysis between weather factors and traffic flow is carried out,and the correlation strength between traffic flow and weather factors is studied,which lays the foundation for the traffic flow prediction that integrates weather factors.Then,use the Pearson correlation coefficient to analyze the time and space correlation of the traffic flow data under the expressway and the ordinary road,and analyze the impact of different weather factors on the traffic flow;use the Spearman correlation coefficient to analyze all the influencing variables Systematic mining is carried out on the relationship between different weather factors and the relationship between traffic flow,and the two-tailed verification of the data matrix is combined with sample testing,which provides a theoretical basis for the study of traffic flow prediction that integrates weather factors.Finally,considering the correlation strength between weather factors and traffic flow,the weather factors are integrated into the process of traffic flow prediction.In the Tensorflow deep learning platform,a support vector regression model,a long and short-term memory network,and a K-nearest neighbor algorithm are established respectively.Traffic flow forecasting model.Use the British M3 highway and A3 ordinary highway data for model training and verification,and analyze the traffic flow data of different time granularities,study the accuracy of traffic flow prediction under different time granularities,and analyze the weather-free factors and the integrated weather factors Compare and analyze the traffic flow prediction results under.The result proves that the traffic flow prediction model based on support vector regression based on 15-minute traffic flow data,which integrates weather factors,performs best.This thesis takes the research of traffic flow characteristics under severe weather conditions as the main topic,deeply analyzes the traffic flow characteristics of different grades of roads under different rainfall conditions,studies the correlation strength of the influence of different weather factors on traffic flow,and constructs a weather factor integration Traffic flow prediction model.The research results can provide important theoretical support for traffic management and control,traffic planning and design,and the development of digital traffic.