The Traffic Characteristics And Capacity Analysis Metohod For Diverge Influence Arear On Multi-Lane Freeways

The traffic demand of freeway in China is continuously increasing.The proportion of six-lane and eight-lane segments in the newly built freeway has been rising year by year.Many trunk routes with saturated traffic have been expanded to six-lane or eight-lane freeways.In the freeway diverge segment and its upstream section of a multi-lane freeway,vehicles frequently change lanes and interweave while various types of vehicles are mixed.The traffic characteristics are more complex than the basic segments and the diverge segment of four-lane freeway.With the influence of downstream ramps and the connected local road,the diverge segment is prone to be congested and is usually the bottleneck of freeways.It is necessary to conduct in-depth research on the influence,traff-ic characteristics,and capacity analysis method of the diverge segment of multi-lane freeway in order to provide theoretical basis and modeling technology for freeway planning,design,construction,and operation.For the analysis of multi-lane freeway diverge segment with complex traffic characteristics and variable operating conditions,the microscopic simulation experiment method has significant advantages.It could be applied under flexible conditions in engineering practice.Also,the development of microscopic simulation models would make certain contributions to traffic operation research.As a research tool,the microscopic traffic simulation system NGTSS(New Generation Traffic Simulation System)has been developed.The paper defines the concept of the multi-lane freeway diverge influence area and concludes the features of the multi-lane freeway diverge segment.Through collecting and analyzing the actual operation data of multi-lane freeway,the traffic characteristics of the multi-lane freeway diverge segment are proposed.The behavioral characteristics of the lane-changing process are analyzed.The average proportions of traffic flow on each lane at different locations upstream the exit ramp are calculated.The average proportions of every vehicle type on each lane of different sections are calculated.Hypothesis testing method was used to fit the features of the speed and the headway data of each section of the diverge influence area.According to the distributions of the lane-changing behaviors and the layout of the traffic signs,the specific scope of the diverge influence area is defined.The key influencing factors of traffic operation in the freeway diverge segment during the planning,design and operation phases are analyzed.The requirements of the microscopic simulation tools in the diverge influence area study are analyzed.The framework of the microscopic simulation system NGTSS are constructed.The study requirements of motion control model,the vehicles' interaction driving behavior model,and the lane changing decision and process models are specified.The structure and correlations of main functional modules and the operation mechanism of the simulation system are designed.The update logic and process sequence of the traffic environment,vehicles' status and drivers' decision variables of each frame of the simulation are deliberately set.The methodology of simulating the delay caused by drivers' reaction time and setting multiple driving modes under different circumstances are proposed.The basic model of the system operation including the road facility model,the motion simulation and control model,the PID control model for the vehicle's direction control simulation and the multiple levels of route decision model are established.The methodology for simulating the interaction between vehicles in the traffic simulation model of the multi-lane freeway diverge influence area is proposed.The algorithms to recognize vehicles related to the subject vehicle and to calculate the time to collision(TTC)for each pair of vehicles are provided.Interaction behaviors of vehicles on the same lane are classified into three scenarios including stable car-following,queueing and collision avoidance.The driving behavior decision models for all three scenarios are established respectively.By simulating the delay caused by reaction time and drivers' memories of the sequent decision process,the existing Psycho-physical car-following model is improved,and a car-following model is established under the first scenario as above.A simulation model of vehicles' parking at the end of a queue and start-up process under congestion is established.The process is divided into three steps as deceleration,queueing and start-up.The speed control simulation methods for all those steps are proposed respectively.A simulation model of collision avoidance behaviors of vehicles is constructed.The indicators of five types of collisions are defined and corresponding strategies are proposed.The simulation models of collision type recognition,decision-making process and safe state discrimination are provided.The modeling framework for lane-changing decision process is established.The decision-making process is divided into multiple levels including target lane selection,gap acceptance,target gap search and final lane-changing execution.At each level,utility values of different choices are compared according to the traffic environment and the drivers' preferences.The calculation method of speed and orientation control during lane-changing process is given.The model to simulate the gaming behavior between the subject lane-changing vehicle and the vehicle which is following it immediately on the target lane under congested conditions is established.The cost matrix of the game theory is given,and different driving strategies are proposed.The proposed lane-changing decision and process model could be applied to analyze the disturbance of diverging behavior on the traffic flow of diverge influence area on multi-lane freeways.The data preprocess method for NGSIM database applied to microscopic simulation calibration is provided.The methodology framework to calibrate the car-following model using microscopic trajectory data and macroscopic traffic flow data is constructed,and the calibrated parameters are given.A statistical analysis method for the control parameters of the lane-changing model is proposed and the calibration result is discussed.The framework and steps to apply the microscopic simulation model in the analysis of the capacity of multi-lane freeway diverge influence area are proposed.The capacity and the standards for different LOS of the diverge influence area for typical facilities in China are obtained through microscopic simulation experiments.The maximum volume in the diverge influence area corresponding to the service levels and different ratios of exiting traffic is calculated.Applying the simulation model to analyze the traffic characteristics under the congested condition,the data analysis method for the development and dissipation process of congestion has been proposed.The average delay of diverge influence area under different flow rates and ratios of exiting traffic has been calculated.