Research on Hard Shoulder Running and Traffic Control Strategies for Accident Sections in Freeway Mixed Traffic Flow

Traffic accidents often lead to severe road congestion. In this study, a three-lane cellular automaton model is developed to simulate traffic flow on highways involving autonomous vehicles under accident conditions. Simulation experiments were carried out to compare the effects of hard shoulder running and HSR combined with lane-changing guidance on mitigating accident-induced congestion. The results show that traffic accidents can cause prolonged and extensive congestion. Opening the hard shoulder significantly alleviates congestion, although a congestion wave tends to form near the closure point of the shoulder. The combination of HSR and lane-changing guidance performs slightly better than HSR alone in reducing overall congestion, with a particularly noticeable effect on dissipating congestion waves in the accident-affected lane. Furthermore, traffic congestion intensifies as the inflow rate increases, eventually reaching a saturation point under medium to high inflow conditions. Increasing the penetration rate of autonomous vehicles helps reduce congestion around accident-affected areas. The combined strategy of HSR and lane-changing guidance is most effective at low to medium penetration rates. When all vehicles are autonomous, both HSR and the combined strategy can reduce congestion to a level comparable to normal accident-free conditions.