When it comes to autonomous vehicles (AVs), there may be more questions today than there are answers.
For example, when will autonomous cars be fully ubiquitous? There are AVs on the road today, but early predictions that they would take over the roads by 20301 are quickly starting to seem unrealistic. In fact, ubiquity may actually be several decades2 away.
Another pressing question: Are consumers ready for autonomous vehicles? While AVs promise to make the streets safer for pedestrians and reduce traffic, some consumers seem unsure about these vehicles altogether.
One study found 45% of new car buyers are hesitant3 to use connected services in vehicles (from navigation to entertainment) due to concerns about privacy. Another found 73% are fearful4 of fully self-driving vehicles.
Altogether, these questions expose a larger uncertainty: what does the future of AVs truly look like? While no one can predict the future, we turned to three industry experts steeped in the data and technology around AVs to understand the trends they’re seeing in the industry today—and the implications for the future.
Meet the Experts:
- Carène Kamel, Business Development Technical Specialist, dSPACE: Kamel helps car manufacturers and suppliers test their vehicles for safety and viability—before they hit the road.
- Jonathan Matus, Founder and CEO @ Zendrive: Matus has a background in mobile devices, and his company Zendrive makes roads safer using mobile sensor data and analytics to improve driver behavior worldwide.
- Prescott Watson, Principal Investor @ Maniv Mobility: Watson is always looking at trends to help his company make smart investments in AV companies.
WIRED: It seems like AV development will be slower than many in the industry predicted. Can you talk a little bit about how you see AVs developing over the next few years?
MATUS: I think we’re going to see a long “teenage” period where autonomous vehicles will be a subset of vehicles on the road.
Consider that the buying cycle for vehicles is about 11 years — and it’s actually becoming longer. If cars were being sold right now with full autonomy, then it would take about 11 years, optimistically, to completely replace the set of vehicles on the road. But of course, these vehicles are not available yet. And those vehicles, even when they do become available, won’t be equally distributed. They’re going to be first distributed in markets like Arizona and California where the weather is perfect and where autonomous vehicle companies are active.
WIRED: What kinds of advancements are needed for AVs before we can realize full autonomy?
WATSON: We’ve seen a lot of investment in sensors, application, and compute, which is great. But two areas I think that we’ve under-invested in are networking and storage. On the networking side, there’s enormous investments in the network outside of the vehicle, like 5G. But we need to invest in questions like, how do you improve the network within the vehicle to get data-intensive applications working, and how do you manage where data gets stored, in or off vehicle?
On the storage side, there’s going to have to be an immense amount of storage in the vehicle, and it’s not simple storage. It’s going to be a heterogeneous type of storage. You’re going to have to adapt the “hot vs cold” data management mentality from the IT world to better fit all the parameters that apply to driving-related data.
KAMEL: We need to effectively plan how these autonomous vehicles are going to coexist on the road, and we’re starting to use simulation technology to help manufacturers test the safety of their vehicles. So we provide a simulation environment, a virtual environment, that represents the world as we know it. It virtualizes the different roads that we drive on and the various traffic participants you can simulate — pretty much any kind of complex traffic scenario that you can encounter in the real life. This way, you have a wide range of scenarios that you can put your autonomous driving functions through to be able to make sure that it’s safe.
WIRED: How will data privacy and security play a role in AV development?
KAMEL: There are different answers to this question depending on where in the world you are. In the U.S., we’re seeing regulatory bodies reacting to the automotive industry. They aren’t trying to regulate everything right away. Instead, they are looking to what the industry is doing on its own to self-regulate. You’re seeing suppliers and car manufacturers joining forces and beginning to define those standards and protocols. Meanwhile, you’re seeing more regulations in the EU as far as AV testing on the road. Here in the U.S, the regulations are coming from the industry, while in Europe there are more regulations coming from the government.
MATUS: I think privacy is the least of the thorny issues that need to be resolved. It’s an important one, but I think it’s one where regulators can require pretty clear user consents and explicit user confirmation.
I think safety issues and traffic control issues are going to determine whether AVs are adopted by the public.
WIRED: What do you see as a more pressing regulatory challenge than privacy and security?
MATUS: Insurance. Together with the shift from normal human drivers to autonomy, there is a shift from personal car ownership, to shared ownership. There are a lot of questions about that, because the current insurance model, and current risk mitigation model is focusing on the car owner — not even the car driver, but the owner – as the central focus for risk mitigation, risk pricing, and claim transaction. But all of this can be very different if you don’t own the vehicle, and if someone is providing this as a service.
WIRED: Do you think we’ll see full autonomy before 2050, or will we see a different version of autonomy altogether?
MATUS: I think full autonomy will come before 2050. But I don’t think it will be equally distributed—there’s going to be pockets where, let’s say, a county, or a city, makes it mandatory, or offers very, very strong financial incentives for using autonomy rather than driving your own vehicle.
WATSON: For me, it’s not a question of when are we going to get robo-taxis to take us everywhere. It’s more, what’s the first application that we’ve automated enough for it to really make a difference?
For example, one of the reasons there’s been so much interest in the is that you don’t have to automate everything about a truck for it to actually improve the way the business works. If you can automate the easiest part of the drive, which is highway driving, and then the drivers actually sleep or do other productive things during the 98% of the miles that are driven on a highway in a relatively controlled environment, you can totally change the economics of the trucking industry. It doesn’t mean the truck is level five autonomy, but actually for the purposes of generating a substantially safer, cleaner and more efficient and profitable business, that level of automation is enough.
This content is produced by WIRED Brand Lab in collaboration with Western Digital Corporation.
- When will self-driving cars take over?
- Self-Driving Cars Won’t Be Ubiquitous for Decades
- What’s Driving the Connected Car
- Three in Four Americans Remain Afraid of Fully Self-Driving Vehicles