From the perspective of driver’s visual characteristics, the potential collision conflicts in car-following and lane-changing driving behaviors are analyzed. On this basis, a bi-directional visual angle car-following model considering the collision sensitivity is proposed. The effect of collision sensitivity coefficient, the proportion of backward-looking visual angle, and many types of vehicles on the dynamics and safety of the newly proposed model is explored. A linear local stability analysis was carried out by perturbation analysis, and the critical stability condition of the improved model was deduced. The evolution of small perturbation during the car-following and lane-changing process is numerically simulated. The same conclusion as the linear stability analysis is obtained: The driver observes different types of vehicles in the vehicle queue through the bi-directional visual angle as feedback stimuli, so as to optimize the velocity and perceive potential collision conflicts, which has significant practical significance for improving the stability of the traffic system and alleviating traffic congestion. Besides, the safety of the new model is evaluated by the improved visual time-to-collision (TTC) index. The results show that the newly introduced factors can effectively reduce the potential collision risk in multiple scenarios.
Read full abstract