About auction strategies for intersection management when human-driven and autonomous vehicles coexist

Autonomous vehicles are appearing in our streets, and will soon populate our transportation infrastructures, which must be equipped with appropriate sensors and actuators in order to manage vehicles in a fruitful way. Besides the infrastructures, appropriate algorithms must be defined in order to coordinate the vehicles and to enable them to exploit the resources in a fair yet effective way. In the immediate future, autonomous vehicles must coexist human-driven vehicles, and this transitory scenario poses several challenges in coordinating both kinds to exploit street resources. One of these resources, whose management is quite challenging, is represented by intersections: vehicles come and aim at passing the intersection, often as soon as possible, but they must compete with other vehicles having the same aim. A possible approach that has been used in literature to this problem uses auction based mechanisms. In this paper, we place ourselves in the above-mentioned transitory scenario in which both human-driven and autonomous vehicles will compete to cross intersections, and we investigate the effectiveness of auction-based mechanism to coordinate vehicles at intersections. We devise some simple auction policies, and assume vehicle coordination strategies that are suitable also for human drivers. Our results lead us to believe that, under these assumptions, simple auction mechanisms do not introduce advantages for what concern traveling times as they do in the case of exclusively autonomous vehicles.