Comprehensive Evaluation Of Drivers’ Risky Driving Behaviors Of Dangerous Freight Vehicles And Research On The Spatio-temporal Characteristics Of Trajectory Data

In recent years,there have been many traffic accidents of dangerous goods on roads in China,which have brought great losses to people’s life safety and economic development.Relevant government departments and scholars in related fields pay great attention to the safety of dangerous goods transportation,and now they pay more attention to the prevention of accidents from the original post-emergency treatment.Therefore,based on the alarm data of 13 types of risky driving behaviors generated in the process of dangerous goods vehicle transportation obtained by Shaanxi Road Transport Safety Monitoring Platform and the trajectory data of GPS terminal when generating risk alarm,this thesis analyzes the advantages and disadvantages of TOPSIS and grey correlation methods and their applicability in the thesis.An improved comprehensive evaluation model of hazardous vehicle driving behavior based on TOPSIS-grey correlation method was established.On this basis,various factors such as high-risk driving behaviors and behaviors prone to traffic accidents are comprehensively considered.Lane departure,fatigue driving and forward collision risk driving alarm track data are selected.After preprocessing the track data,the time distribution results of the three types of risk driving alarm track data are studied.DBSCAN clustering algorithm is used to mine risky driving alarm hot spots,and Arc GIS tool is used to realize the visualization of the alarm hot spots.The results:(1)Through actual comparison and verify,it can be concluded that the TOPSIS-grey correlation method constructed in this thesis can combine the advantages of TOPSIS and grey correlation method,and scientifically and effectively comprehensively rank the incidence of 13 types of risky driving behaviors from high to low from two aspects of position and shape.(2)The high occurrence time points of lane departure,fatigue driving and forward collision driving alarm track data were concentrated in the early morning,late night and the peak traffic period;Alarm hot spots are mostly concentrated in tunnels,sharp turns,steep slopes and other places,and the three types of alarm behavior have different causative factors,and the clustering results have their own characteristics,which are analyzed in detail in this thesis.At the same time,according to the results of the alarming analysis of risky driving behavior,suggestions on the safe driving supervision of dangerous vehicles are put forward from three aspects: strengthening the supervision work of government departments,improving the safety supervision efficiency of enterprises,and improving the construction of transportation infrastructure.