Have driverless cars stalled?
Or are they just taking a circuitous route to becoming reality? Nicole Kobie reveals what’s holding up automotive automation
Or are they just taking a more circuitous route to becoming reality? Nicole Kobie talks to the experts to find out what’s holding up automotive automation.
Driverless cars have been just around the corner for more than a decade. We were supposed to take our hands off the wheel from 2020, according to car industry leaders Ford and Nissan
– and yet here we are, still requiring driving licences.
Some progress has been made. Owners of Teslas can flip a switch into Autopilot mode, with beta users given advanced features that extend the car’s automation to “full self driving” mode – although documents filed by the company reveal the setting still requires human driving. Google’s Waymo is picking up taxi fares using cars without a human safety driver at the wheel; instead, engineers sit at the ready remotely to help the AI if it faces a situation it can’t figure out. And Apple has plans to unveil its own driverless car in 2024.
Despite this momentum, car makers admit that progress hasn’t met their own deadlines, whilst Waymo’s CEO John Krafcik has conceded that driverless tech would always have “some constraints”. The company said at the end of last year it would offer entirely driverless rides in a select area of Phoenix, but the cars are remotely connected to engineers in case of challenges and some will still require safety drivers.
Perhaps, finally, the hype bubble has been popped and reality has set in. “I think we’re starting to see a correction happening at the moment,” said Jack Stilgoe, associate professor at University College London and author of Who’s Driving
Innovation? “The paradoxical thing is that the earlier a technology is, the more hype there is. The more work people actually do into the technology, the more they realise how hard it is to bring it into the real world.”
He added: “Obviously, we shouldn’t listen only to what the companies say. They’re trying to sell stuff. But that isn’t to say that the technology won’t have enormous value in the years to come.”
Technical challenges
Creating such a technology while AI remains in its relatively early years is no small feat. That means it remains likely that fully driverless technologies are years in the future, although there’s no question that it’s possible for cars to navigate streets by using cameras for eyes and a mix of sensors to gather even more data. The challenge remains to ensure machine learning can understand the images and data passed back to it by those sensors and make decisions based on that information. For that, systems are trained in real-world scenarios – that is, they’re driven a lot – and also in simulations.
Driverless cars use three main vision systems: radar, cameras and Lidar. The latter stands for “light detection and ranging”, and bounces pulses of light off objects to determine their size, shape and, if they’re moving, speed. Each of those vision systems have their own weaknesses – radar, for example, doesn’t work as well in bad weather – and the subsequent AI-based interpretation can be impacted by everything from precipitation to spray-painted signs.
If you see a “Stop” sign covered in graffiti, you understand braking is still required; if snow covers lanes, you go slowly and follow the car
“The more work people actually do into the tech, the more they realise how hard it is to bring it into the real world”
“I don’t believe people understand that most cars already have a level of autonomy – we are just gradually increasing this”
ahead. For truly driverless cars, AI systems need the same mental flexibility. But, to take just one example, Waymo’s cars have spent much of their time in the sunny states of Arizona and California. Amid reports that the cars didn’t work well in heavy rain, Waymo sent a few down to Florida in 2019 for the wet season to help train them for driving in adverse conditions.
Weather isn’t the only challenge. Trials might be zooming ahead in the US, but the road networks in many American cities are simpler than British cities, according to Camilla Fowler, head of automation at TRL. “The infrastructure we have actually adds a lot of complexity to the driving task,” she said.
How and why
Automation is already here, but there’s more to driverless technology than waiting for engineers to work out the kinks of fully driverless cars. The challenge is far more subtle.
“The question is quite often posed in terms of ‘when’ we will see driverless cars,” said Stilgoe. “I think the real question is actually where and in what form.”
If we can’t solve all the problems through engineering, there are other solutions: we could limit where the cars are used, cap their speeds, rework roads or pick and choose the elements of automation that do work.
For example, some cars will be geofenced – limited in where they can travel because the developers know to a precise degree that the system works well enough in a specific area. Or they could be limited to special lanes. “The technology could in some cases need dedicated infrastructure, whether that’s a dedicated lane, or smart infrastructure like smart traffic lights,” Stilgoe added. Indeed, researchers at Princeton University have developed radar reflectors to alert automated cars to the presence of pedestrians and cyclists on roads.
Another example is automation only being allowed on motorways, rather than more complex side roads that also feature pedestrians and cyclists, while “automated lane keeping systems” are set to be approved in the UK this year.
That iterative approach to automated features is already happening with computer-aided parking and braking, for example, and is one way that driverless technologies will creep into our lives. “I don’t believe people understand that most cars actually already have a level of autonomy – we are just gradually increasing this,” said Bani Anvari, research leader of the Intelligent Mobility Group at University College London.
That idea is expressed in the six levels of autonomy: from level zero with none whatsoever to level five with the ability to drive anywhere and everywhere without human intervention. Perhaps rather than be entirely driverless, cars will be partially automated (level two) or have conditional automation (level three), which means the driver may be asked to retake the wheel. The levels are helpful for regulators. Tesla was forced to admit in paperwork filed to the California Department of Motor Vehicles that what it calls full self-driving mode is actually a level-two car requiring human supervision.
While lane-changing assistance and keeping a car creeping forward during a traffic jam may not be the stuff of sci-fi films, such features could help reduce fatigue and exhaustion, and are more likely to be in your next car than fully level-five automation, which may never be technically possible.
Note the “may” in that sentence. Plenty of tech leaders believe that level five isn’t just possible, but likely soon. Elon Musk has made it clear that he believes Teslas will be the first level-five driverless cars, suggesting that software updates will make it possible for an existing Tesla to navigate roads wholly unaided. That said, he claimed in 2020 that Tesla would have level-five cars by the end of that year, and now it’s 2021 and his
“If these cars aren’t deemed to be safe by the wider public, people will keep their own hands on the wheel”
company is filing papers with the DMV about level two.
TRL’s Fowler predicts that full automation will happen first in controlled circumstances where it’s actually useful, such as around campuses or down residential streets. “It depends on what society needs,” Fowler said. “What’s the impact going to be on people’s wellbeing, safety, network capacity – benefits need to be realised in order for this tech to be realised, I think.”
Indeed, wherever driverless is used, the point is to solve a problem, not just roll out technology for the sake of it. “I want governments to be clear on what they want to see from this technology,” said Stilgoe. “If they reduce congestion, are sustainable, are safer than alternatives, if they benefit people who have previously lost out on mobility technologies – those things a city could be particularly keen on. Rather than just starting with technological possibilities, you start with what people need.”
Regulate the brains
The tech industry tends to lament regulation, but driverless cars is one market that demands a robust testing and approval process. After all, any “dumb” car currently on the road is required to undergo thorough testing before being allowed to be sold. Now that we’re cramming brains into vehicles, we need to test those brains.
But we’re only just figuring out how to assess algorithms and AI, and that’s before they’re put into machines at high speeds. So how do we know when driverless cars are safe? “What does the manufacturer need to do to demonstrate that all the different use cases and scenarios have been considered that the vehicle could come across within its operational design domain?” to use Fowler’s phraseology. “How do you validate, how do you approve?”
While safety is key – after all, no one will use driverless cars if they aren’t safe – Fowler notes that regulators must ensure the approvals process doesn’t create a barrier to development and innovation. “We need to be enabling it, rather than holding it back,” Fowler said.
Building trust
The point of that regulation is to build trust: if these cars aren’t deemed to be safe by the wider public, people will keep their own hands on the wheel. There have been a few driverless car accidents, notably the two deaths in Teslas with Autopilot mode enabled and the death of Elaine Herzberg in a collision with an Uber car in Arizona. So far, such incidents haven’t yet set back research in this area, although Uber did sell off its autonomous car research division.
Instead, what could prove the more serious hurdle is the hype itself, by sparking a disconnect between what the wider public believes driverless cars can do at the moment and in the future. “I find it worrying that people feel the technology can do more than it can,” said Fowler. “That misconception of what your vehicle is capable of, and not fully understanding the limitations, could be a potential issue.”
And University College London’s Anvari argues we need to understand how the systems work to ever fully trust them; we need to set aside the hype for honesty. “If you have an autonomous minibus, if the people sitting within it don’t understand what this technology is observing and how it takes actions, they won’t gain trust,” Anvari explained. “If they can see this technology can see what they see, and it’s reacting the way they would react, it will help gain trust.”
Anvari continued: “If the whole human side of things is not solved, the trust of users is not solved, you will not be successful in rolling out the technology.”