Cooperative Driving at Intersections Through Agent-Based Argumentation

Stefano Mariani, Dario Ferrari, Franco Zambonelli

Abstract: In a future of self-driving and connected vehicles, cooperative driving will be the key to guarantee that not only isolated vehicles can hit the road safely on their own, but that the collective of vehicles displays efficient and safe behaviours. Intersection crossing is arguably the most challenging problem for cooperative driving, as vehicles need to coordinate their relative movements while avoiding collisions and optimising intersection throughput. In this paper, we propose a multi-agent based approach exploiting computational argumentation to coordinate vehicles at intersections: vehicles approaching an intersection, represented by agents, argue about their right of way, while an arbitration process resolves conflicting arguments (i.e., leading to vehicle collisions) by applying a configurable conflicts resolution policy and suggesting alternative routes to vehicles. Extensive simulation results show that – in most situations – the argumentation-based approach enables increasing the overall throughput at intersections while decreasing vehicles’ delay.