In the spring of 2020, as the COVID-19 crisis raged, several Chinese technology companies quietly executed a curious pivot in their self-driving-car strategies. The multinational AI specialist Baidu, in a June press conference, disclosed that it had been sharing its autonomous vehicle technology with a wide variety of partner companies — free of charge, for several months — to help fight the pandemic.
According to Chinese media reports, several other companies also collaborated on ways to repurpose their automation technology. The results have been kicking in ever since. One Baidu partner, previously a maker of agricultural robots, shipped out 30 autonomous disinfection vehicles to hospitals in multiple cities. Other self-driving vehicles were refitted to deliver meals, medicine, and water into quarantined areas. Still others patrolled streets and sidewalks, scrubbing and cleaning, and reminding residents, via loudspeaker, to wear their masks.
These ideas have been catching on. In the U.S. last spring, self-driving shuttles delivered COVID-19 tests between facilities in Florida, with the aim of minimizing person-to-person transmission of the virus. The self-driving car service Cruise, a General Motors subsidiary, customized several of its autonomous passenger vehicles to transport food to the needy in the Bay Area.
Taskin Padir (above), Northeastern University engineering professor
“Pretty much every technology company is experiencing some kind of slowdown.”
If there’s a twinge of cognitive dissonance here, you can blame the marketers. In the U.S., especially, the term “self-driving car” conjures visions of shiny electric sedans drifting through nighttime cityscapes in beautifully lit television commercials. But as these pandemic-related developments attest, automated vehicle technology has the potential to serve entirely different purposes. It’s a grim yet reliable fact that major upheavals like wars and pandemics tend to accelerate technological development. The COVID-19 pandemic may change the ultimate trajectory of self-driving cars.
After a promising start in the 21st century imagination, the space-age vision of the self-driving car has plateaued in the last several years. The fully autonomous vehicle — what the industry calls Level Five technology — has proven elusive. Despite sophisticated object-recognition and AI technology, computer-driven cars simply cannot account for the million-and-one things that can happen on the road. Instead, we’ve settled for incremental upgrades like traffic-jam assist and cars that can parallel park themselves.
But the promises keep coming. The online sphere is packed with cheerful conjecture on what far-future, privately-owned robot cars might look like. For instance, Volkswagen’s ID Buzz promises a revolutionary new version of the old Microbus, with a driver’s seat that swivels around backward so you can hold a rolling business meeting. Forward-looking designers are even looking to modify the partially automated Tesla Semi tractor-trailer to create an AI-powered RV.
Thanks to COVID-19, however, the self-driving car industry is hitting the same roadblocks that are slowing down everyone else. Adverse economic conditions are taking a toll, forcing many major players to change plans or delay development of automated vehicle initiatives, says Northeastern University engineering professor Taskin Padir, who specializes in automation technology and directs the Robotics and Intelligent Vehicles Research Laboratory at the university.
“Pretty much every technology company is experiencing some kind of slowdown,” Padir says.
In the months after the crisis hit, Ford Motor Co. postponed the commercial deployment of its autonomous vehicle for a year. Waymo, the heralded self-driving unit of Google parent company Alphabet Inc., temporarily suspended several of its testing operations. It’s been even worse for the smaller players. The New York Times reported that at least one U.S. autonomous vehicle startup is going out of business, another is for sale, and four more are laying off employees.
“Some have had the resources to keep going, but others just won’t make it,” Padir says.
Another major challenge is the fact that, for several years now, the development of self-driving technology has been bound up with car-sharing and ride-sharing services. The pre-COVID road map for the industry assumed that most people would have their first autonomous vehicle experience in a rental or robo-taxi. But these companies are struggling, too, as social distancing and sanitation concerns scare people away. General Motors recently scuttled its car-sharing service Maven, and ridesharing giants like Uber and Lyft — each working on its own autonomous projects — have been in crisis control mode.
The readiness factor
But the biggest obstacle to the widespread use of fully-automated vehicles has nothing to do with COVID-19. Instead, it’s the core technology itself.
“We’re just not there yet,” says Jason Torchinsky, senior editor at the car-centered website Jalopnik and author of Robot, Take the Wheel: The Road to Autonomous Cars and the Lost Art of Driving. To be truly viable, widely deployed transportation, says Torchinsky, self-driving cars would need to reach the technical stage known as Level Five, or complete autonomy. They’d have to drive themselves safely in any road condition, any type of weather, any location in the world.
A quick primer: Levels One through Three refer to driver-assist technology, in which the car can handle some things — lane changes, braking, acceleration — but the driver must be able to take over at any instant. Level Four is considered “high automation,” in which the car can operate without human input, but only in designated areas, like an interstate highway.
True Level Five technology would be so advanced that most passengers would find the experience underwhelming, says Andy Schaudt, program director for the Center for Automated Vehicle Systems at the Virginia Tech Transportation Institute.
“You’re going to get in the vehicle, it’s going to drive the speed limit and obey all the traffic laws, and then it’s going to drop you off at point B,” predicts Schaudt. “It’s not going to be like a roller coaster ride or something that’s just out of this world.”
Right now, Torchinsky says, the majority of autonomous cars are at Level Two or Level Three technology — with a few advanced systems, and operating in controlled areas, such as closed-off testing facilities or tiny commercial plazas.
“We just can’t trust yet that these vehicles are going to be able to handle anything and everything,” he says. “And if they’re supposed to be safer than humans — if that’s the whole point of this thing — then yeah, it’s still a long way off.”
Northeastern’s Padir says that in the best-case scenario, current driverless technology can handle around 90 percent of driving situations reliably. But that’s not nearly good enough. For people to trust driverless cars — to actually just hop in and read a book or close their eyes — Padir thinks the technology needs to be as reliable as commercial aviation: 99.999 percent or better.
“And that last mile is always the hardest part,” Padir says.
The middle distance
This is where COVID-19 comes in, because right now we could really use some other options for moving people and provisions around.
The global lockdowns of the spring made that need feel acute. As the virus surged, viral anxieties kept people away from most conventional modes of travel. Between February and August 2020, commercial airline travel dropped as much as 91 percent in the United States alone, while long-distance rail travel plummeted 90 percent. In U.S. cities in October, monthly public transport ridership was 65 percent lower than it had been a year earlier. International numbers have been equally alarming.
Industry watchers say autonomous vehicle technology can provide some potentially effective solutions to our COVID-era transportation anxieties — if we acknowledge their limitations, adjust our expectations, and start thinking creatively.
James Sayer, director of the University of Michigan Transportation Research Institute, says automated technologies such as home delivery of groceries, medications, and other necessities will come online years before mass ownership of fully-automated vehicles for personal use.
Transporting large numbers of passengers in high-speed automated vehicles is still years away. “Maybe decades,” says Torchinsky. But we can aim for more modest solutions, like low-speed robotic transport carts that can minimize person-to-person interactions.
“This is one of the leading vectors of development right now, and it’s probably where autonomous vehicles could have the most use in future scenarios,” Torchinsky says. “You order online from the grocery store or the drugstore, and the robot car brings everything to your home. That would be a real benefit, possibly even more than moving people around.”
And with careful development, autonomous vehicles can potentially fill other needs during pandemics, like the sanitation bots in China. With low speeds and appropriate distancing, we could even start moving people around.
“We may see expanded deployment of small-scale people-mover-type operations,” Sayer says, “perhaps low-speed shuttles operating in closed areas, like in a senior-citizen community.”
Who needs a ride?
Some technologists believe these crisis-driven shifts should compel the transportation industry to re-think the more distant future of autonomous cars, and to build scenarios around equity more than individual convenience.
“When considering the people who might use automated transportation, I see two very different groups,” Sayer says. “Those who could purchase a fully-automated vehicle are those who are affluent, who likely are able to work from home anyway.”
But front-line and essential workers, he says, are less likely to be able to afford a self-driving car, and more likely to need their benefits.
“These are the people who are at the greatest risk of exposure and who are often in cities that rely on public transportation,” he says. “These are people for whom mass transit is their only real solution. It becomes a public health response.”
Torchinsky imagines a far-term future where self-driving vehicles fill in for crowded public transportation options.
“There could be fleets of them,” Torchinsky says. “It would definitely be a lot better than the options we have now, in terms of the virus. You’re not cramming people in a bus, which is just going be an incubator for that stuff.”
Indeed, researchers are already exploring new sterilization technologies. Schaudt says his transportation research group has tested many sanitation strategies in ride-sharing and car-sharing systems — air filtration systems, disinfectants, and infrared and UV light. If these technologies grow in parallel with the automation technology, he says, companies could develop vehicles that sanitize themselves between trips, like those self-cleaning restrooms in airplanes and transit stations.
One crisis at a time
In the COVID-19 era, Sayer says, it’s time to repurpose autonomous vehicle technology for tasks where it will do the most immediate good.
“Designers, manufacturers, and researchers could be using this time to ask themselves important questions,” he says. “What opportunities are there for my industry to help in the recovery, to address the realities my customers face day-to-day?”
Northeastern’s Padir agrees that autonomous vehicle developers have an unexpected opportunity, right now, to reshape the future of transportation. “We need to innovate now, today,” he says, “to be ready for the next challenge that society will face.”