The impact of selfish driving behavior of autonomous vehicles on mixed traffic flow

With the development of automatic control and artificial intelligence technology, autonomous vehicles (AVs) are becoming more and more intelligent. However, this individual intelligence means that the driving behavior of AVs can be more aggressive and selfish. To investigate the impact of the selfish driving behavior of AVs on the characteristics of mixed traffic flow, this paper proposes a cellular automata model that considers the selfish driving behavior of AVs. First, we analyze the vehicle types‌ in the current mixed traffic flow and develop a universal safety distance model. Then, based on this, a cellular automata model is proposed. This model can not only describe the longitudinal car-following and lateral lane-changing behavior of human-driven vehicles and AVs, but also characterize the selfish degree in the driving behavior of AVs. Finally, simulation experiments are designed to analyze the impact of the selfish driving behavior of AVs on the performance indicators of mixed traffic flow. The results show that: (1) The increase in the penetration rate of AVs has a positive effect on the average velocity, traffic stability, and throughput of mixed traffic flow; (2) On the whole, the selfish lane-changing behavior of AVs has a relatively small impact on the average velocity of mixed traffic flow. The impact becomes smaller with the increase in traffic density. (3) In most scenarios, the selfish car-following behavior of AVs has an improving effect on the average velocity and throughput of mixed traffic flow. In summary, the relevant research results can provide theoretical support for the design of micro-behavioral control schemes for AVs.