The Vision (Part 2): Tackling the tough issues
In part two of our tech talk, Continental experts Andree Hohm and Björn Filzek get down to business: how do we ensure that AD cars drive safely? And will people be ready to entrust them with their lives?
Now let’s get to the tricky questions. How do you guarantee that an automated car is driving reliably? After all, electronics can fail...
Filzek: True, but this is not a new challenge. There are tried and tested procedures to make electronics failsafe. Take for example electronic brake systems: any critical failure is obviously intolerable. To ensure this there are elaborate development processes in place, extensive product testing is applied and comprehensive monitoring measures are implemented. We need to tailor those to the new use cases.
How can you be sure automated cars drive safely – in case of a malfunction or other danger?
Filzek: This has to be addressed by a three-step approach: firstly, we find ways to minimize the risk of a malfunction. But nobody will be able to rule them out 100 %. This leads us to step two: if a malfunction occurs, rare as this may be, we will make sure the system recognizes that immediately – and takes protective action. The third scenario deals with dangers from third parties: for instance, another driver makes a mistake, a deer suddenly jumps in front of your car. In those cases, sensors and connectivity will often give the vehicle a timely warning: accidents through lane changes for instance will be minimized if one vehicle can tell the other: “Hey watch out, I’m changing the lane!” In the rare cases where an impact cannot be avoided, there is still the whole spectrum of passive safety measures available like airbags for example. These layers make sure that automated cars will drive safely!
What’s the technology behind these safety measures?
Filzek: One simple, but effective principle is redundancy: if one sensor fails, you have a second in place to do the job. This is combined with high-end monitoring mechanisms: sensitive points in the system will be under constant surveillance so that a malfunction is immediately recognized.
“We will have answers for every scenario”
What exactly does an automated car do, once it realizes there’s an imminent safety threat?
Filzek: It will immediately get to a safe state. What that maneuver looks like, depends on the traffic situation the car is in. If it’s in a traffic jam, the vehicle may simply stop. If it’s going fast on a highway, it has possibly to be able to make it to the hard shoulder, even from the left lane. If there is no hard shoulder, the car will know that and drive accordingly. It is a cascading approach where we have to identify tailored answers for every scenario.
How real is the risk of cyber attacks?
Hohm: Again, cyber attacks are not a new threat. A prominent case has made the headlines last year in which hackers successfully attacked a car that wasn’t even fully automated. Obviously, the more connected a device becomes, the more potential entry points for attacks there are. But at the same time the countermeasures reach unprecedented efficiency – for example with new levels of encryption and firewalls. Just take a look at where mankind is already using connected systems in highly sensitive contexts: just think of air traffic or military technology and how fatal a hacker attack might be there. Yet, engineers have succeeded in making these systems sufficiently safe. The same will happen for automated driving.
How do consumers feel about these issues? Are they enthusiastic about AD – or worried?
Hohm: Our studies show that there is a great level of enthusiasm for this technology – people are excited by it, they see the advantages. At the same time there are diffuse concerns about the topics you just mentioned. This is a gap we need to bridge. We need to get to a point where using an automated car feels like taking the train: you don’t even think about the tracks ahead or what is going on in the locomotive – because the system has proven to you and so to say to the society that this transportation device is trustworthy.
“The solution is called transparency”
How can this be achieved? How can you make sure people see the automated vehicle as a helpful friend and not a foe?
Hohm: First of all, it’s a good thing that AD won’t happen overnight. People will slowly get used to advancing levels of automation. This will build trust.
Do you see a danger that even small steps of automation irritate the driver?
Hohm: No, since there is a clear solution. It’s called transparency. I compared automated driving to riding a train – but it can’t be like this overnight. You can’t just disconnect a driver today from what is happening around him. While AD is still new and drivers are getting used to it they need absolute transparency about what their car is doing and what it bases its decisions on. It is a fascinating technological challenge to create a human-machine-interface that delivers this kind of transparency in an intuitive way.
Will smartphones be a source of inspiration for car interfaces? They’re already pretty user-friendly…
Hohm: There is a whole community of great minds from science and industry working on this, fueling each other with ideas. Our experts are part of this community. We are looking at a multitude of solutions and sources of inspiration – also from the field of consumer electronics.
How will I operate my car’s interface in the future? Will I talk to it? Will I use gestures? Can you already detect a tendency?
Hohm: Not yet. We are currently working on providing the full toolbox of possible functions to provide to the car makers. At the end it is up to them to choose the ideal configuration for their vehicles out of solutions provided by us.
Will it be difficult to reach a common standard? It might get confusing, if each car has a different interface…
Filzek: I’m confident that they will all converge to a kind of natural optimum. Think of the smartphone: In its early days, the swiping function on touchscreen displays was completely new and unparalleled. Now almost every phone has it. The same will happen in cars.
So we will have uniform AD vehicles?
Hohm: The foundations will be standardized: if you know how to drive a car in New York, you also know how to drive a rented car in, say, Bangladesh. But there will also be room for individual solutions.
Filzek: Car makers can still design the look and feel of signals and controls to their liking. And there will be solutions to tailor interfaces to individual customers – so that they provide the ideal experience, no matter what state the driver is in.
In the third and last part of our tech talk, the experts make an urgent plea towards governments.