An entirely new view of all aspects of a vehicle is essential to design moving forward.
Future generations of vehicles will age like any other electronic or mechanical devices, but they also will need to adapt, grow, and change in unexpected ways over time to avoid being hacked, rendered obsolete, or otherwise compromised.
This adds a whole new set of challenges never seen before in automotive development, and OEMs are working feverishly to bring system architectures up to speed for driver assistance, safety, security and increasingly autonomous driving features.
At the outset, that requires a strategy for how these vehicles can be secured in place, in transit, and when downloading over-the-air updates. This is the same problem that the medical industry has been wrestling with in recent years, with a few exceptions.
“[The medical industry has] the exact same problem, except that they have had connected devices out there for years in the home, in the hospital, and these things are just as open to being attacked by Mirai as everything else is,” said Robert Bates, chief safety officer in Mentor Graphics’ Embedded Software Division. “But in the car, while we’ve got well-publicized attacks, there aren’t a lot of vectors that are legitimately available to be attacked today because cars aren’t really all that connected. It might have a 4G or LTE radio for ONSTAR or emergency communication-type applications, but there’s a reason the hacking demonstrations occur with someone close to the car. You have to be close to the car in most cases to really impact it.”
There is an opportunity to take what has been learned here and apply security to new systems from the ground up. “If we don’t, five years from now we’re going to be sitting here, and we’re not going to be talking about necessarily cyberterrorists taking over our cars. But we’re certainly going to be talking about cyber criminals holding our cars ransom to the automakers for millions and millions of dollars,” he said.
Cheapest communication available
Before the security can be applied, the architecture of the vehicle must be understood, and how the communication within the vehicle is structured. That includes how the vehicle-to-vehicle, and vehicle-to-infrastructure, communications are enabled.
“Today, in GM’s ONSTAR, for example, the pipe coming into the car is say, 4G, and the pipe inside the car is Bluetooth and WiFi,” said Richard Barrett, senior director, product marketing at Cypress. “Sometimes the pipe outside of the car is WiFi, as well. It’s much cheaper to do over-the-air [updates] off of your home media server or some form of connection to WiFi, and that’s going to continue to drive WiFi technology. It’s doing it today at faster and faster WiFi speeds for OTA for downloading media, and so forth. You’ll use whatever network is the cheapest.”
Not surprisingly, the choice comes down to the cost model of what is cheapest and most efficient, he said. “If WiFi is cheaper for you, you’ll use that, assuming you’re not in a mobility situation where you can’t get a link. If 4G or 5G is cheaper, that’s what you’ll use.”
In this way, a blend of technologies will occur, but whether that kills other potential technologies remains to be seen, he noted. “There’s been talk about dedicated short-range communications (DSRC) IEEE 802.11p for years for many car-to-car applications. Are these going to start to atrophy the value of that? It already has, to some extent. Other technologies like ultra-wideband went by the wayside. IEEE 802.11ad WiGig looks interesting, but it’s expensive. Some newer and older technologies that are proven sort of squeeze out some of the paths that people are looking at and funding right now.”
Ron DiGiuseppe, senior strategic marketing manager at Synopsys, agreed that many technical trends or decisions will be market-based. But he suggested a lot of those will also be regulatory and government driven mandates. “For example, vehicle-to-vehicle communications, or vehicle-to-infrastructure communication, and in the context of managing traffic 5G would be ideal for that given the very low latency, very robust, high reliability. But we also have the U.S. Government Department of Transportation looking at the IEEE 802.11p for the vehicle-to-vehicle communication. There’s a bit of industry alignment that needs to occur. Is vehicle-to-vehicle going to be 5G, or the DSRC 802.11p? I don’t think that’s been resolved yet.”
Barrett believes this will be impacted greatly by safety considerations. “If it’s life-critical, the government wants to get involved. If you’re passing information back from the car that’s nice to have, then you can use other networks.”
The National Highway Transportation Safety Administration’s (NHTSA) rules have not yet been released, but they are expected before the Obama Administration exits. The currently target is the IEEE 802.11p DSRC radio protocol, as it relates to collision avoidance systems, which will be dedicated systems.
Regardless of the details of the first U.S. regulations, this is going to become a worldwide standard, Bates noted. “The consortiums are going to have to come together and work this out.”
At the same time, he noted there is a pretty compelling argument that vehicle-to-infrastructure communication needs to be 5G. “These cars are going to be on the road for 20 or 30 years, and the only thing that I know that works that is 10 or 15 years old in technology is that I can bring up an old EDGE or analog cell phone and it still connects to the network. I can’t be guaranteed about that for any kind of emerging technology that’s coming out. Especially when we’re talking about vehicle-to-infrastructure, 5G is going to be the only realistic worldwide solution, and that can be backed up with something like 802.11p, in an individual country. I just don’t see anything else being viable. Industry will drive this like functional safety. These things are all now driven by standards bodies with representation across the world.”
If 5G prevails, there are some specific opportunities for the automotive ecosystem.
“One of the interesting ones is that with the progression to 5G, in addition to the latency and overall connectivity, that allows a carrier or service provider to [essentially] turn their business into a micro-services business,” said Charlene Marini, vice president of competitive strategy at ARM. “Right now, it’s very hard for a carrier to have multiple business models, so they service the mobile handset market. And traditionally for automotive today or M2M (machine to machine), they have MVNOs (mobile virtual network operators) or others that fill the subscriber base for that segment of the market. With 5G and [related technologies] on the back-end side of 5G, it’s much easier for them to balloon their service models, and almost do a service model per individual. When you think about autonomous cars or semi-autonomous cars—people in the vehicle and the driver—there are huge opportunities to offer new services when someone is not fully engaged in the driving experience. And again, there are opportunities for the passengers independent of the timeframe for autonomous vehicles. This is a whole opportunity for service providers and application providers that today can really only provide services in the home or mobile context.”
Another very apparent opportunity is ride-sharing, Marini said. She pointed to the recently-launched Lynk & Co., a Swedish-Chinese company that allows a vehicle owner to push a button to say they want to share their car for a specific period of time.
Regarding opportunities in the ecosystem, DiGiuseppe said it goes top to bottom through the whole supply chain. “When it comes to big data and ridesharing, there is tremendous change for the OEMs — probably more change now than in the last 100 years —but also the whole supply chain, including the Tier 1s, the Tier 2s, and all the software suppliers through the whole chain. This is affecting the whole ecosystem.”
The technology disruptions are readily apparent, but the ecosystem impacts might not be as obvious initially. “In automotive, we have a very established ecosystem today, and the introduction of Mobileye was the first disruption in a long time,” said Samer Hijazi, senior design engineering architect in the IP Group at Cadence. “We are going to see a lot of that going forward.”
He said this is comparable to early laptops, where the operating system was included, then more was paid to put applications on the laptop. “In the same way, our cars will become a platform, and the data traffic is going to be made money on, and services will be provided to us through this connection. It will be a whole new ecosystem of software that does not exist today.”
But cars are very different from laptops. So who will be responsible for all these applications?
Bates suggested that while everyone is talking about 2019 for fully autonomous, he’ll believe it when he sees it. Insurers will be the big question. “Right now, it’s pretty easy. You’re responsible. Even if you go mostly full autonomous in a Tesla, you’re supposed to be there and alert and basically driving, so it’s kind of simple. If I get into a car that doesn’t have a steering wheel or brake pedal or anything like that, and something happens, how could I be responsible for that? All of that has to be worked out, and frankly, I don’t think that any of this is going to roll out in a big way until that is all agreed upon.”
Car as living platform
When it comes down to it, a new view of the vehicle is underway. ARM’s Marini noted that through the supply chain everyone needs to increase their software capacity and understanding. “Very real in this transition to 5G, connected cars, and increased security, is that the car is a living platform. One of the things about mobile phones, which still isn’t true in IoT in general as well as automotive, is renewability — separation and renewability. You separate areas that are not going to be secure, and you ensure that you can update both the secure and the non-secure side so if there is an attack, if there is somehow penetration, you can renew that device. So this evolving, living platform concept is really what we need to get to in order to get to something that can achieve everything we are talking about doing.”
DiGiuseppe agreed, adding that because there are so many millions of lines of code in today’s cars, software integrity and checking software buffer overflows, and all the safety and security of the software — not just the hardware — will become tremendously important.”
And even that may only be the beginning.
Will 5G Enable Connected Cars?
New standards are being considered for automobiles, but there are still a myriad of issues to be resolved.
What Can Go Wrong In Automotive
Experts at the Table, part 1: Security, diagnostics, standards and the future of autonomous vehicles.
Auto Security And Technology Questions Persist
Fallout could be slower adoption of autonomous vehicles as ecosystem proceeds with caution.
Waiting For 5G Technology
New wireless standard will significantly speed up communication, but dealing with mmWave technology isn’t going to be simple.