Combining Coordination Strategies for Autonomous Vehicles in Intersections Networks

Different strategies have been proposed to date to coordinate autonomous, connected vehicles in crossing an isolated intersection, and their performances analysed. However, few works assess what happens in a network of intersections each managed with a different strategy, a situation likely to occur during the transitory phase when not all intersections in a network can (or are worth to) be equipped with a dedicated computational infrastructure to apply a specific strategy. Accordingly, we analyse the impact of combining different strategies in an intersections network. The key result we achieve is that, even in the presence of a few intersections managed by strategies different from currently mainstream ones (such as traffic lights or precedence), significant improvements to global traffic are measured.