Simulating collective decision-making for autonomous vehicles coordination enabled by vehicular networks: A computational social choice perspective

Abstract Recent advancements in Intelligent Transportation Systems have endowed vehicles with autonomous driving capabilities as well as V2X communication. Vehicular networks allow vehicles to augment their perception range beyond Line of Sight conditions, enabling novel strategies for effective conflict resolution in challenging scenarios (e.g. platooning) making use of (Multi-Agent System [MAS]) coordination mechanisms (e.g. voting). Simulation-based evaluation of coordination mechanisms should consider real-world constraints, such as kinematics and communication constraints (e.g. packet loss). To this effect, this paper presents a simulation framework that unifies vehicular, communications and MAS simulators in order to test more effectively and realistically these mechanisms. The framework is benchmarked using a lane-merging and platoon voting scenario. Results show the scalability of the framework and the capability to model coordination mechanisms in next to real-world conditions.