A Right-of-way Negotiation System Using Inter-Vehicle Communication on a Road with a Mix of Autonomous and Human-Driven Vehicles

摘要

The Japanese government aims to introduce completely autonomous vehicles by 2030. Autonomous vehicles are expected to reduce the number of accidents and traffic jams. However, incomplete exchange of yielding intention between human-driven vehicles and computer-controlled autonomous vehicles, which will run on same roads, may lead to accidents or traffic jams. This paper proposes a system that enables vehicles to accurately transmit their yielding intention to other vehicles via inter-vehicle communication, irrespective of whether the vehicles concerned are driven by a human or a computer. We have identified seven patterns of yielding situations between vehicles and developed a system that enables safe yielding in each pattern. We have implemented the proposed system in Raspberry Pi computers and mounted them on radio control cars in order to evaluate operations of the system. The computer on each car recognizes the road width, nearby vehicles and their types by analyzing images captured by the camera mounted on each car. The interface between the computer and the driver is implemented using a smartphone that is connected via Bluetooth to the Raspberry Pi computer on the car. We have conducted experiments to measure time needed to make a yielding decision in a variety of driving conditions and evaluated the accuracy of these decisions.