Research On Traffic Characteristics And Safety Effect Of Electric Bicycles In Urban Streets

With the growth of national economy,the urbanization process is advancing rapidly in China.The expansion of cities,acceleration of motorization and rapid growth of travel demand of urban residents make urban road traffic in China face a severe challenge.With the expansion of cities and the rises of personal incomes in China,more and more trip makers are shifting from bicycles to electric bicycles.Although they provide users with convenient,flexible and affordable mobility,electric bicycles have raised some issues and concerns regarding their safety impacts.The dissertation supported by Projects of International Cooperation and Exchange of the National Natural Science Foundation of China,focused on the efficiency and safety of residents' travel.The key issues on non-motorized vehicles were discussed in the dissertation,including the characteristics of non-motorized traffic flow,the capacity of non-motorized lanes,the level of service of non-motorized lanes,riders' risky behaviors and safety evaluation on non-motorized vehicles.A method was proposed to estimate the capacity of mid-block bicycle lanes and the bicycle equivalent units based on the composite headway probability density distribution of two-wheeled vehicles in mixed traffic flow.The composite headway distribution models were developed to identify the probability density distribution of the unconstrained and constrained headways of the electric bikes(e-bikes),electric scooters(e-scooters)and bicycles in mid-block bicycle lanes with mixed traffic flow.The composite headway model was applied to estimate the capacity of mid-block bicycle lanes with mixed traffic flow composed of the e-bikes,e-scooters and bicycles and the bicycle equivalent unit for the e-bikes and e-scooters.The study identified the factors that influenced the comfort perception of e-bike,e-scooter and bicycle riders in mid-block bicycle lanes on urban streets and estimated the bicycle level of service(BLOS)of a mid-block bicycle lane with mixed two-wheeled traffic.The factors that significantly affected the comfort perception of the cyclists included the age of the cyclists,the type of two-wheeled vehicles,the volume of two-wheeled vehicles,the width of mid-block bicycle lanes,the proportions of e-bikes and e-scooters in non-motorized vehicles,the physical separation between motorized,non-motorized and pedestrian lanes,the slope of non-motorized lanes,the roadside access points and the roadside land use.Ordered probit models were developed to quantitatively evaluate the impacts of different contributing factors on the comfort perception of the riders of e-bikes,e-scooters and bicycles.With the comfort perception models,a procedure was developed and insights were gained to help transportation professionals estimate the BLOS of a mid-block bicycle lane with mixed two-wheeled traffic.The conflicts between motorized and non-motorized vehicles were divided into sixteen conflict types in term of the right-of-way of road users and theoretical traffic conflict points.The causes of each type of the conflicts were discussed.It was found that the predominant cause was that the drivers of motorized vehicles did not yield to the right-of-way of non-motorized vehicles.The time to collision and conflict rate were compared between electric bicycles/regular bicycles and motorized vehicles.A metod was proposed to calculate the comprehensive traffic conflict rate in signilzed intersections.The study compared the risky behaviors of e-bike,e-scooter and bicycle riders as they were crossing signalized intersections and identified the factors that influenced the risky behaviors of e-bike,e-scooter and bicycle riders.Four different types of risky behaviors were identified,including red-light running,stop beyond the stop line,riding in motorized lanes and riding against traffic.The factors that significantly affected the risky behaviors of the cyclists included the type of two-wheeled vehicles,the gender and age of the cyclists,the group size,the peak period,the type of pedestrian signal,the width of signalized intersections,the volume of non-motorized vehicles and the conflict volume of motorized vehicles.The study identified the influencing factors to traffic conflicts between non-motorized and motorized vehicles at signalized intersections and those between two-wheeled vehicles in mid-block bicycle lanes.Three types of models were compared,including the fixed-effect,the random-effect and the random-parameter generalized linear regressions models.It was found that the random-parameter models provided the best goodness-of-fit to field data of signalzed intersections.The factors that significantly affected the conflict frequency in signalized intersections included the volume of non-motorized and motorized vehicles,the peak period,the channelization,the width of non-motoized lanes,the physical separation between motorized and non-motorized lanes and the proportion of e-scooters in two-wheeled vehicles.The fixed-effect model provided the best goodness-of-fit to field data of mid-block bicycle lanes.The factors that significantly affected the conflict frequency in mid-block bicycle lanes included the volume of electric bicycles and regular bicycles,the average speed of electric bicycles,the average speed difference between electric bicycles and regular bicycles and the width of non-motorized lanes.