PARIS -- Renault Group is partnering with Sanef, a French highway operator, to develop a communications network between autonomous vehicles and the road infrastructure, starting with a pilot project to help self-driving cars navigate through construction and toll areas.
While self-driving cars can navigate static situations such as highway travel by using cameras, radar and lidar to “read” road signs and markings, less predictable, more congested challenges such as work areas and toll zones strain the limits of current technology.
“The challenge for the autonomous car is to make these choices,” said Edouard Fischer, the chief technical officer of Sanef, which is a subsidiary of Abertis, a Spanish company that operates toll-collection sites. “You have to integrate a lot of information in real time.”
When a vehicle approaches a toll zone, it needs to decide which lanes are closed and which method of payment to use. Then, it needs to be able to pass through a narrow toll booth at up to 30 kph, in the case of automatic payments. Adding to the difficulty, multilane toll exchanges do not have ground markings, so cameras cannot transmit information about lanes.
Under the system being tested, roadside transmitters send information about which lanes are appropriate for autonomous vehicles, and the car will begin to slow. A high-definition map of the exchange creates “virtual” lanes for the car to follow.
Testing is currently taking place on French highways using Renault’s autonomous vehicle prototypes.
The problems are similar for work zones or events such as winter de-icing, Fischer said. Data can be transmitted by trucks positioned near the projects, or by the de-icing vehicles or sensors embedded in the road.