Self-driving technology is mired in a slump.
Argo AI, one of the leading companies developing robotaxis, closed in October. Some companies have been battered in public markets, bringing the specter of a delisting from stock exchanges. Two substantial crashes renewed concerns about overall safety in 2022.
Some companies see a path through this Trough of Disillusionment — those with the right “combination of capabilities, capital and technology” will “go and succeed,” Aurora Innovation CEO and co-founder Chris Urmson said last week at the Automotive News Congress in Detroit.
But other companies are seeking an exit ramp. Automakers such as Ford Motor Co. are looking to recoup some of their massive investments in self-driving by utilizing some of the underlying technologies for conditional automated driving.
Ford is among the automakers more focused on commercializing what automotive engineers call Level 3 automation, rather than trying to jump to Level 4, in which humans have no role in the driving process. CEO Jim Farley told the Automotive News Congress there is a lot of value to be unlocked.
“People are willing to pay a lot of money for that, and the stress release of taking your hands off the wheel,” he said. “And the time back? Unbelievable. We’ll never see pricing power like when Level 3 gets democratized.”
Level 3 automated systems may offer human motorists that magic moment that has proved so elusive with self-driving technology: the ability to remove their hands from the wheel and their attention from the road, freeing them to ostensibly do things such as read a book or watch YouTube.
Crossing the threshold from Level 2 to Level 3 constitutes a major leap. Level 3 systems do more than take active control of driving, like many current driver-assistance systems already in production. They go one step further: When enabled for highway driving, the systems themselves take on responsibility for vehicle operations, though humans might be needed to retake control should complications arise.
But before the full hype of Level 3 arrives with the marketing of these new systems, beware that they bring their own complications and a new round of unanswered questions.
Among them: What happens if a situation arises and a human driver does not respond to prompts to retake control? Might cars need redundant equipment to ensure cars can navigate safely to a stop? Can those cars pull to the side of the road or do they stop in their lane of travel? At highway speeds, does the a car sopped in a lane constitute a safety hazard?
It’s easy to foresee a cascading series of questions.
How much time should a motorist be given to retake control when prompted in a regular circumstance? In an emergency? What constitutes a true emergency and what scenarios should cars be expected to handle? There are no industry standards right now.
Is it reasonable to believe a motorist engrossed in a book or YouTube or a nap might be thrust into driving at a moment’s notice and have enough understanding of the road and traffic environments to evade a problem in such an emergency?
Are law enforcement officers trained to understand the nuances of these technologies when investigating a crash? Do they possess the proper equipment and technical background to discern who or what controlled a vehicle at the time of a collision?
The focus may have downshifted from full self-driving to Level 3 automation. But it is wishful thinking to believe the scope of the challenge has gotten simpler. At each level, automation brings with it distinct complexities. Solving any of them is not a task in the rearview mirror, but a responsibility very much remaining on the road ahead.