A Costs ‘Reality Check’ is Coming To AV Manufacturers

Years of hype suggest that driverless cars are right around the corner, inching toward deployment with just a few more miles of road tests.

General Motors is one of the automakers leading the charge with plans to bring its Super Cruise technology to all Cadillac models by 2020, followed by other GM brands in the years after. Waymo also hopes to make a big push in 2020. Despite the ongoing promise to deliver AVs so quickly, this might be one case where fantasy has greatly surpassed reality.

“We believe there will be a reality check and all that hype will come down a little bit and people will understand that it will simply take time,” said Patrick Hupperich, president and CEO of FEV North America, a vehicle development services provider. “The benefits of going autonomous, they are compelling in certain use cases but, for mass deployment, we are a little bit skeptical.”

Hupperich said that development costs remain the largest hurdle to building driverless cars for the mainstream market. A big part of that comes from LiDAR, which is one of the most expensive components currently used in self-driving vehicles. Millions (if not billions) of venture capital dollars have been poured into LiDAR start-ups, some of which are focused on making cheaper modules. This may lead to the technological and economical breakthroughs that carmakers are hoping for but Hupperich said that the level of choice presents other challenges.

“It’s difficult with the choices that you have and the products and the technology that is being developed at a rapid pace,” he said. “To keep up with that and really find out what is the technology that has the biggest potential in the future.”

While Hupperich believes that mass deployment will happen eventually, he doesn’t seem to buy into the anticipation that surrounds the industry. He expressed concern about infrastructure, communication protocols and other technologies that are not yet ready for other tech devices, let alone driverless cars. “There’s hype on the West Coast, one start-up after another,” he added. “We believe there will be some consolidation in the next two to three years. This is clear. The fittest will survive and come out of this stronger than before.”

Signs of consolidation began with GM’s acquisition of Cruise Automation which, in turn, acquired a LiDAR sensor developer, Strobe. Uber quickly followed suit and purchased Otto, the now-defunct AV tech company focused on building driverless trucks. Aptiv made its own sizable acquisition when it picked up nuTonomy. Now GM and Cruise have teamed up with Honda to further AV development.

“There will be businesses that emerge around this that we [could not anticipate] today,” said Hupperich. “The OEM might turn into a supplier, into a mobility service company. The whole make and use of data that is generated by these vehicles – even if they’re not connected via V2X or V2V, there are endless opportunities. That’s why partnerships and consolidations are key to compete, cooperate and co-integrate.”

Survival of the fittest

Those who survive will be the companies that figure out how to monetize their investments but many obstacles remain. Hupperich identified connectivity as one of the biggest issues that needs to be addressed.

“We need vehicles to really be connected and not just GM vehicles being connected to GM vehicles,” he said, favoring a model where all vehicles communicate openly, regardless of make or model. “There’s a big discussion for 802.11p or 5G. That’s, of course, very important that the industry agrees on certain technologies people can build upon. The challenge of 5G is [that] it’s not ready. From an infrastructure standpoint, you could use a lot of what exists, whereas 802.11p, big investments need to be made for the infrastructure to be built. In a geofenced area, there’s probably no problem that the vehicle can communicate beautifully with its environment.”

May Mobility is among the start-ups that have tested their AV technologies in a low-speed, highly geofenced environment. The company currently relies on its own infrastructure (deployed throughout its testing area in Detroit, Michigan) to maintain the necessary level of connectivity.

Long-term, Hupperich theorized that a combination of connectivity technologies might be needed for self-driving vehicles to ensure that bandwidth does not become an issue. “If you have mass deployment and every vehicle is a beacon and uses a cellular network, and everybody streams videos in their cars, in terms of data congestion, is 5G going to be able to handle it?” Hupperich questioned. “Nobody really knows but it will take a few years just to roll out. Maybe it will be a shared use of 5G for non-safety-related functions, so you can focus just on exchanging data that is not safety critical, and then you use the other platform for safety-related functions.”


Leave a comment

Your email address will not be published. Required fields are marked *