Characteristics Of Traffic Oscillation Based On Driver Behavior

With the acceleration of urban development,the problem of urban road congestion has gradually become more prominent,especially in the rush hours.Traffic flow theory analyzes road traffic flow and studies the internal laws of traffic flow.It also reveals the formation and development mechanism of traffic congestion which can help alleviate the current status of traffic congestion and provide scientific guidance for traffic management and control.As a part of traffic flow theory,traffic oscillation has been widely studied by scholars in recent years.The formation and development of traffic oscillation seriously reduce the efficiency of road traffic and increases vehicle fuel consumption.It is also a major hidden danger of road traffic safety.Driver behavior is an important part of the research on car following behavior.There are significant differences in the behaviors of different types of drivers in the car following process.It can provide a clearer understanding of the car following behavior from the perspective of the driver.Incorporating driver factors into the car following model can improve the accuracy of the car following model and describe the car following behavior better.This paper first organizes the drivers to carry out the car-following test and obtains the required car-following data,and then uses the moving average method to denoise the data and analyzes the driver's speed difference period under different car following speeds by wavelet transform.The difference of the speed difference period in the driver's car following process is analyzed by significance test.Finally the drivers are divided by the cluster analysis method.Based on the traditional IDM model,this paper analyzes the correlation of parameters in the following behavior and studies the relationship between the headway,space headway and speed difference,then establishes an improved IDM car following model.Finally,we calibrate the model by genetic algorithm and verify the improved model.Extracting traffic oscillation data,we analyze the concave growth pattern of the standard deviation of the driver's speed in the fleet with the increase of the vehicle number.Then we use the improved car following model to reproduce the concave growth of the standard deviation of the speed.Traffic hysteresis during the car following process and the differences in traffic hysteresis of different types of drivers are analyzed.Finally,traffic hysteresis is verified by the improved IDM car following model.