In partnership with Audi. Could driverless cars become the new normal? Advances in the technology and ethics surrounding automated vehicles are pushing us ever closer to transforming our streets. What was once only possible in science fiction could soon be an observable everyday occurrence. We look at how driverless car systems have improved over the last decade.
Driverless vehicles are already being tested on our roads but it’s going to take a little longer for people to catch up to the idea that they’re as safe as they would be in vehicles driven by humans. The software-development team at Audi is working hard to achieve a fully automated future of driving – and to put people’s minds at rest.
As far as visions of future mobility go, the image of a driverless car is perhaps the most referenced and discussed of all. In many ways automated vehicles are symbolic of our attitudes towards technology and how it is redrawing our daily habits. Yet much of the debate is still focused on the hypothetical scenarios that driverless cars could create in our cities and on our streets if (or when) people stop being behind the wheel altogether. Much less weight is being placed on the fact that artificial intelligence has been in the process of entering our vehicles for years – and its effects can be felt in the present too.“Over the years we have made steps based on different technical platforms,” says Roland Pfänder, manager of software development at Audi Electronics Venture GmbH. He is the leader of one of many teams working on creating driverless solutions within Audi. “But overall I would say that we’ve been working on these systems for the past 10 years. It’s a different kind of development – conceptual work – but ours is a race that didn’t start yesterday.”
From speed control to assisted parking, as well as ways of ensuring that a car doesn’t stray into another lane when driving down a highway, plenty of functions have already been developed so as to support driving rather than substituting it. Introducing these features in widely available car models is a progressive step-by-step approach that not only follows the natural evolution of technology development but is handy when it comes to allowing motorists to get used to a whole different way of approaching being in their cars. “The best way to get people convinced about the technology is to give them a chance to experience it,” says Pfänder. Those who find it hard to believe they could ever relax in their seats while their car takes care of their commute may well find their vehicle’s ability to slowly ferry them through a long traffic jam (a feature that the developers at Audi are planning for their cars to include within the near future) is a potent persuasive tool.
“When we discuss the ability of our systems there is a classification of levels,” says Pfänder. “Where we are now is level three, which is automated driving systems, and level five is for the customer to have no necessity to interact with the car for driving.” The commercial availability of fully automated systems may still lie in the future but it is towards that goal that Pfänder’s team is striving. The task at hand is huge: what we need is to construct an algorithm-based software, combined with artificial intelligence, that not only replicates a human’s ability to learn and decide but also improves on it.
“Driving is a very specific task in terms of cognitive capability – a complex one, of course – so what we’re doing is trying to break it down into sub-tasks,” says Pfänder. “One of them, for example, is environmental perception, which means understanding what’s going on in the street around the car.”
In many ways the availability of data – and the ability for it to be shared between different vehicles – will become essential to creating systems that can control (and adapt) to traffic, truly transforming the way our streets are run. With a clean slate and regulatory frameworks still to be defined, the process that will result in a driverless future is very much all to play for: the careful practice of those who taking a staggered approach can lead the way.
Philosophy professor Patrick Lin is the director of the Ethics and Emerging Sciences Group at California Polytechnic State University. Over the course of a career in robot ethics that has spanned more than a decade he has witnessed and helped shape the discourse around driverless cars. Here he weighs in on our own dilemmas.
Q There are obvious practical advantages to the use of driverless cars. But how willing are drivers to step into an automated vehicle?
A Novelty is unsustainable so we tend to get over new technologies quickly. But the first steps are often the trickiest. Automated cars will need to quickly prove that they’re much safer than human drivers. We need to be able to trust the technology as well as trust that technology developers are taking safety and ethics seriously – for instance, not viewing possible accidents as part of a cost-benefit calculation. Otherwise social acceptance may be a much longer, rockier path.
Q Do you think it’s possible that driverless cars could end up actually simplifying a person’s ethical choices?
A Well, ethics is not always simple; if it is then you’re doing it wrong. But that’s the point of both policies and automation: to streamline decision-making by following a set procedure. If you have a good policy then you don’t have to keep wrestling with the same issues every time. The danger, though, is in oversimplifying our choices. Some scenarios introduce factors that a policy might not have considered so it’s not enough to defer to policy; you have to put in the work to account for those new variables. Likewise, automated driving could be helpful in quickly determining how much distance you should leave between your car and another motorist, or how fast you should drive given road conditions, or what the safest action would be if a deer popped up in front of you. But we need to be clear on the limits of that automation and ensure that they have been developed thoughtfully. It would be unethical and irresponsible not to conduct that basic due diligence.
This is the fifth instalment in an eight-part series that considers the car of the future. From urbanists to engineers, we speak to the experts who can decode the changes in the world of mobility.