Road Freight Truck Operating Status Identification And Risk Recognition Based On Trajectory Data

In recent years,the number of traffic accidents involving trucks has remained high,and the safety of freight transportation has become increasingly severe.The rapid development of big data technology has provided new data sources and research methods for freight research.Trajectory data contains rich vehicle feature information and potential operating rules,and it is necessary to further extract knowledge represented by data and explore hidden operating feature information in travel data,which can provide decision-making basis for real-time monitoring and safety assessment of vehicles for operational enterprises.Based on the truck trajectory data,this paper identifies the truck stopping locations and recognizes the truck operating mode according to the identification results.It extracts the potential freight characteristics in the truck trajectory data,constructs a risk assessment index system based on the operating features,quantifies the level of truck operating risk using fuzzy set theory,and finally evaluates the truck operating risk.The specific research work is as follows:A recognition model for freight truck travel patterns is built.An algorithm for extracting truck parking points based on mobile records is designed using the trajectory data of freight trucks.A method for identifying parking purposes by integrating freight truck travel trajectory data is proposed.A feature set that includes parking time features,stay trajectory features,and nearby point of interest features is constructed as input variables for the recognition model.The XGBoosting algorithm is employed to identify on-the-way temporary parking,loading and unloading parking,and rest parking.Based on the recognition results,a travel chain for freight trucks is constructed.Clustering algorithm is used for clustering based on six main features of the travel chain,including the average number of stopping points for each travel chain,the average dwell time for each stopping point,the average travel time for each travel,the distance traveled for each travel chain,and the proportion of selecting similar trajectories,which can classify freight truck travel patterns into four categories:long-distance random route travel patterns,short-distance random route travel patterns,short-distance fixed route travel patterns,and circular travel patterns.The truck operation risk assessment index system was constructed in this study using passive voice and the present perfect tense.After the truck’s travel-related features were identified,six evaluation indicators for truck travel risk were proposed from four aspects:travel speed,travel time,travel parking,and travel distance.The algorithm for obtaining each indicator was described,and the distribution characteristics of the indicators were studied to provide a foundation for subsequent quantitative research on truck operation risk based on the truck’s travel features.The truck operation risk assessment index system was constructed in this study using passive voice and the present perfect tense.After the truck’s travel-related features were identified,six evaluation indicators for truck travel risk were proposed from four aspects:travel speed,travel time,travel parking,and travel distance.The algorithm for obtaining each indicator was described,and the distribution characteristics of the indicators were studied to provide a foundation for subsequent quantitative research on truck operation risk based on the truck’s travel features.