It’s not just technology that holds the key to the autonomous-vehicle revolution. Automakers, suppliers, IT companies and telecommunications service providers are hard at work to create vehicles that will, one day, take the driver out of the equation. But, before that happens, there are non-technical issues, too, that need to be addressed. By Arjen Bongard.
Human factors raise a large number of questions, especially during the current transitional phase on the road to autonomous mobility. Many of these issues are studied by behavioral psychologists such as Marieke Martens.
Martens is director of science at Dutch research organization TNO and also serves as professor intelligent transport systems (ITS) and human factors at the University of Twente in The Netherlands. She specializes in human responses to in-car and on-road smart mobility solutions
In an interview with automotiveIT International at the ITS World Congress in Copenhagen in September, Martens made clear there’s always another side to the technological leaps and bounds that are reported almost daily by the companies working on driverless cars.
Take road congestion: The common wisdom is that more autonomous and shared vehicles will reduce traffic density. But more autonomous cars could actually increase the number of vehicles on the road. Says Martens: “The moment you make things easier, people will start to use it more.”
More miles may also be driven once people realize that they can use the time stuck in traffic productively. “When you can open your laptop and do some work, it will start to feel like regular working hours and not like lost time,” Martens says.
Then there are scenarios that predict people will use autonomous cars more for journeys they would otherwise not have undertaken. Returning home to pick up a forgotten wallet would be a case in point.
While the exact impact of autonomous vehicles on traffic congestion is still unclear, Martens says people have begun to take traffic delays in stride. “The problem still is unexpected congestion, which we really don’t like,” she says. “But predictable traffic jams we apparently don’t mind so much.”
Experts predict that driverless cars will be safer than those piloted by humans, but, before these vehicles hit the road in larger numbers, there are several complex problems that need to be overcome during the transitional phase.
Take the handover to a human driver, which needs to happen when driverless technology decides that’s the safest option in a given situation. Humans have to take control regularly in current, so-called level 2 automated vehicles and they will also have to be ready to take over in level 3 cars.
That will increasingly be a challenge as the technology improves. “The better automotive systems become, the worse people will be at paying attention,” Martens says. “Their attention wanes, they’re no longer adjusting their behavior appropriately and their situational awareness declines.” The result: The time it will take a driver to assume control of a vehicle will be longer.
New technology can help deal with this challenge. Martens says highly automated vehicles will benefit from more and better information about their surroundings. “A lot is predictable these days,” she says, mentioning road construction, the dynamics of traffic buildup, accidents, broken-down vehicles and cueing traffic.
Also, driver monitoring is one way to improve safety and facilitate the handover. Martens says knowing more about the driver is an important tool. With the help of driver monitoring, autonomous technology can predict how long a handover is likely to take. A warning to the driver to get ready can then, for example, be given earlier.
Desire for autonomy
Martens sees discussions about the ethical decisions an autonomous car must make in an accident situation premature. She cites a recent MIT study that asked people how a driverless car should react when a deadly accident involving one or more people is unavoidable. “Once we’re at the point where this becomes relevant, I’ll be very happy,” she says, “because we are not there by a long stretch.”
The safety of today’s autonomous technology should not be measured by miles driven without accidents, Martens says. Rather, the industry should attempt to prove how many accidents have been avoided by the technology. “You have to analyze the data in a totally different way,” she says.
Martens says people probably want driverless cars, even though surveys of potential car buyers in many countries show widespread skepticism. “There’s a big difference between asking people and offering them something real,” she says. Martens argues that different people will have different reasons for wanting a driverless car at a particular moment. But she’s convinced the new form of personal mobility will catch on. “If there’s a clear benefit, it’s affordable and comfortable, everyone will want it.”
Nevertheless, the TNO psychologist expects the current transitional phase on the road toward full autonomy to take a long time. That’s first and foremost because driver-operated vehicles have been the norm for more than 100 years and people have to get used to a totally new form of personal mobility. Says Martens: “People’s confidence has to grow.”