Page 32 - CBT 2018

30
Optimizing intersection control for connected autonomous vehicles
Traffic signals are designed to safely control human-driven vehicles at intersections, but connected autonomous vehicles admit more complex intersection protocols relying on precisely coordinated vehicle maneuvering. An intersection manager uses dedicated short-range wireless communications to reserve trajectories for individual vehicles with specific entry times and speeds. A
natural question is finding the vehicle trajectories that maximize intersection throughput while avoiding collisions. Vehicle trajectories can intersect along
a finite number of conflict points within the intersection, which can be calculated geometrically based on possible turning movements. The time each vehicle will occupy a conflict point is determined by its length and speed. Safety is ensured by preventing overlap in conflict point occupancy times. By encoding the safety constraints as linear equations of conflict point arrival times and integer ordering variables, vehicle throughput can be formulated and solved using mixed integer linear programming. [47]