Why Renault thinks it can succeed in autonomous shuttles where others failed

GUYANCOURT, France -- Renault Group has drawn a clear line between driving assistance for passenger cars and full self-driving for public transportation.

The strategy may seem unusual for an automaker, but group executives say investing in costly and potentially dangerous autonomous vehicle technology for private buyers makes little sense for Renault as a mass-market brand.

On the one side of the line are advanced driver assistance systems (ADAS) such as lane-keeping that still require the driver to pay attention and keep a hand on the wheel, or SAE Level 2. On the other side are true self-driving systems that range from Level 3, where the car takes control under limited circumstance and the driver can be asked to take over at any time, to Level 5, where the self-driving systems function in all conditions.

“To make it simple, Renaut Group intends to work on Level 2 for passenger cars and light-commercial vans,” CTO Gilles Le Borgne said at Renault’s technical center in Guyancourt, France. “But we also want to work on Level 4 to fulfill needs regarding public transport.” 

Renault says it will continue to prepare its cars for self-driving, if conditions eventually allow it, but Le Borgne said the need for bulletproof reliability for passenger vehicles on open roads can make the cost prohibitive. A true self-driving car would need to have no more than one failure per billion hours of operation, he said. For a single car, that cost is born by the owner, but on public transit, it is shared among many users.

Level 4 systems work in what is known as a controlled operational design domain, or ODD, somewhat simplifying the safety issue. That generally refers to a closed loop or route that a bus would travel.

Renault is focusing its Level 4 efforts on public transit – specifically, automated shuttles – because it believes it can succeed where a rash of startups have struggled. 

Among high-profile failures was France’s Nayva, which filed for receivership last year. Others that ran into difficulty include the U.S.'s Optimus Ride (technology acquired by Magna) and Local Motors, and Switzerland's Bestmile (technology acquired by ZF Friedrichshafen). 

Renault’s advantage, said Le Borgne, is that it has expertise in the “hardware” – the platforms, the components and the manufacturing expertise – to support the self-driving technology.

“Renault brings the best of both worlds: High-volume LCV manufacturing plus technology from partners such as Easy Mile and WeRide,” said Jean-Francois Salessy, head of advanced vehicle engineering at Renault Group.

The first autonomous shuttles from Renault Group will be built in partnership with China’s WeRide, in which Renault has an ownership stake. They will be based on the full-electric Renault Master large van, and eventually a new EV architecture from Renault’s Flexis “last mile” venture with Volvo Trucks. Sales are set to start by the end of this decade.

New regulations in France and Germany will actually allow such shuttles on the road, Salessy said. To showcase the technology, Renault and WeRide will make several shuttles available to tennis fans at this month’s French Open. The buses have six lidar units as well as cameras; riders can see what the bus “sees” through overhead screens, one with augmented reality avatars. The buses are limited to 30 kph (19 mph). 

On a test ride around the technical center, the WeRide shuttle behaved generally flawlessly, although it was set to react conservatively to oncoming cars and cross traffic, said Patrick Vergelas, head of autonomous mobility projects at Renault Group, who was along for the ride. Acceleration and stops were smooth and quiet.

Salessy said that autonomous shuttles meet the needs of riders, public authorities and businesses that operate them.

• For riders, they are simple and reliable, with timetables that aren’t affected by unforeseen conditions such as traffic jams.
• For municipalities, they are quiet and clean, and much less expensive than trains or buses.
• And for operators, they offer flexible configurations, and huge savings on personnel. Salessy said that driver salaries can account for two-thirds of the operating costs, but a single human controller can remotely keep track of 10 to 15 autonomous shuttles.

He said the market for autonomous shuttles is potentially thousands per year starting in the 2030s, especially as low-emission zones increase throughout Europe. “Most cities can’t afford light rail or even electric buses, so we see a very significant market for autonomous shuttles,” he said. 

Le Borgne said the first shuttles will be built on a Master van base to get them on the road fast and convince more customers of their viability, especially outside of France. "We have the green light. We are ready to go," he said. "Once you have real use cases, it will be easy to convince others."