Focus: Managing Autonomous Vehicles at Intersections

Since 1970, centralized traffic-control systems have been installed to meet the demand of increasing mobility. Thus, roads can be used much more flexibly. Nevertheless, resources for centralized systems lack scalability, and the maintenance of such systems involves high cost and much effort. The significant and rapidly increasing improvements in communication and information technology, as well as the decreasing prices, are motivating a new look at the handling of the growing roadtraffic problem. The idea is to put more intelligence into the vehicles to benefit from decentralized solutions. However, the problem’s complexity rises with the number of participating vehicles. One promising solution is for each vehicle to put its own computing power into the system, thus helping to meet the growing demand for computing power. As an example, we’ve developed an intersection-management strategy for autonomous vehicles. In this strategy, a vehicle crosses an intersection in coordination with other vehicles, collision-free and following a predetermined strategy-for example, to optimize traffic flow or minimize emissions. This relieves the driver of his or her responsibility within a “sphere of influence” around the intersection where the management is at work. The driver also relinquishes some individuality in favor of