交通震荡传播与驾驶行为关系的仿真研究OA

A mixed traffic flow scenario was constructed using the SUMO simulation platform.Aggressive,stable,and conservative human drivers,along with adaptive cruise control(ACC)vehicles,were defined through a parameterized approach,by which a large dataset of unexpected events and their impact ranges was generated.A LightGBM classification model was then constructed,with road-segment traffic flow parameters and the proportion of different vehicle types in the platoon behind the incident used as input features to predict the impact range of unexpected events.The contribution of each feature to the model's predictions and the underlying mechanisms were revealed using the SHAP method.The results showed that the proportion of aggressive drivers in the following platoon was the most important factor in determining the impact range of an event.The proportion of ACC vehicles exhibited a complex nonlinear effect,and under certain conditions,particularly when the proportion of aggressive drivers was high,it was shown to actually increase the risk of high-impact events.These results confirm that in mixed traffic flows,the microscopic composition of driving behavior,particularly the distribution of risk preferences among drivers in the platoon,is the primary factor determining traffic flow stability.