Space-Efficient Traffic Protocols for Intelligent Crossroads

With the advent of autonomous driving, intelligent crossroads aim to substitute conventional traffic lights by interleaving vehicles crossing in all possible directions. To this end, there must be sufficiently large gaps between vehicles on the different lanes, which needs to be enforced by a traffic protocol taking vehicles' dimensions into account. In particular, existing such protocols are designed for the longest possible vehicle resulting in space-hungry intersections, that require modifications in the infrastructure (in particular, broader roads/lanes), Moreover, these do not allow for vehicles that are exceptionally longer than the ones considered at design time (e.g., extra long trucks, or buses, etc.), In this paper, to overcome this limitation, we present a traffic protocol which handles overlength vehicles as exceptions to compensate for their low probability of occurrence, relaxing space requirements on the intersections. We perform a detailed analysis and comparison showing that the proposed approach leads to high vehicle throughput while keeping intersections small, in particular, as overlength vehicles become longer and less frequent.