Bridging The Gap Between Smart Cities And Autonomous Vehicles

Smart city planners and carmakers are wrestling with similar problems and goals, but they are working at very different paces and often with different technologies — despite the fact that these two worlds will need to be bridged in order to be useful.

Moving traffic optimally in urban areas is critical for reducing energy consumption and accidents, and for moving emergency vehicles through city streets as quickly as possible. How best to achieve these goals is complicated, though. It requires an interleaving of communication, security, and power technology, which needs to be installed in roads, traffic signals, buildings and other structures, with communications from vehicles talking to that infrastructure, as well as to other vehicles.

At the center of this is a contentious debate in North America between the communications standards 802.11p (DSRC) and Cellular Vehicle-to-Everything (C-V2X), which is based on 5G.

Fig. 1: Vehicles communicate with each other and infrastructure in a smart city. Source: Siemens EDA

In November, the U.S. Federal Communications Commission adopted new rules meant to improve automotive safety by reserving the upper 30MHz of the band for Intelligent Transportation System (ITS) services, and designating Cellular V2X as the technology standard for safety-related transportation and vehicular communications. C-V2X uses cellular protocols to provide direct communications between vehicles and obstacles like other vehicles, cyclists, pedestrians, and road workers, and to receive safety information from roadside transmitters.

C-V2X has gained momentum both in the United States and internationally. The FCC had designated Dedicated Short-Range Communications (DSRC) services as the technology standard for ITS services more than 20 years ago, but DSRC has not been meaningfully deployed. As a result, this critical mid-band spectrum has largely been unused for decades. So the FCC has begun transitioning away from DSRC  toward C-V2X, “to hasten the actual deployment of ITS services that will improve automotive safety.”

This is a significant shift because DSRC is incompatible with C-V2X. As with many such standards-related activities, this is not yet a settled matter, and developers are pressing on.

“Autonomy will happen, irrespective of C-V2X, because it’s happening,” said Jim Misener, senior director of product management at Qualcomm. “But C-V2X will considerably enhance the idea of autonomous vehicles, because you have vehicles in close proximity doing message exchange. There is no sensor that can predict the trajectory of a vehicle unless it’s that vehicle communicating its future trajectory. There’s no sensor that can look a little bit at least around intersections except for radios and C-V2X.”

Until now, this development has been vehicle-centric or regulation-centric. “In order for utility to happen on day one, you’re going to have to talk to someone if you have C-V2X in a car,” Misener said. “There will be low market penetration day one. The first utility would be to talk to the city, so we are keenly aware of the need from the car side, but also from the services and value engendered on the road side. The road operator transcends cities, by the way, because road operators are sometimes regional, sometimes at a state level, but the preponderance of roads are in cities, and the pain points happen with traffic engineers in cities so we see the safety coming in first,” he explained.

There also will be efficiencies in terms of cars talking to intersections, via smart traffic signals, whereby the traffic signals can start adapting because they know what the traffic patterns are. Here, a large market penetration is not required. For example, if a car is near the end of a queue, that’s sufficient. The car transmitting the data doesn’t have to be at the exact end of the queue. This starts to obviate the need for roadside sensors, which are quite expensive.

Another example involves emergency services vehicles and first responders. “What if they could themselves dial in with C-V2X green all the way, and do clearance for the entire trajectory — not just in the near field? That’s a combination of the wireless wide area network and short range,” he said. “It would make responses pretty quick, and the benefits happen instantly. That speaks to operational efficiency. A smart city is not a static set of infrastructure pertinences. A smart city is dynamic, so any road user is part of a smart city including pedestrians, and vehicles.”

While V2X and autonomy will develop concurrently, they most likely will have separate schedules for implementation.

“Even for use cases where lidar is used at an intersection, it doesn’t mean it’s for an autonomous Level 4 or Level 5 vehicle,” noted Willard Tu, director of automotive at Xilinx. “It could be for a Level 2 system that says a car is coming from one of your blind spots. In fact, a lot of vehicles could offer that capability, and that wouldn’t be as expensive as Level 4 or 5. On the other hand, in commercial vehicles like trucks, that use case is a lot more complex, and it will necessitate more expensive technology. That also has a different infrastructure, which will be deployed on corridors, but not necessarily in densely-populated areas. A first use case could be depot-to-depot, followed by ‘last-mile’ scenarios, which will be needed in highly urbanized areas. So while these are parallel paths, it would make sense to build different infrastructure for the corridors.”

Then there is the vehicle itself. How V2X concepts play out within the vehicle is still being worked out, although clues are emerging.

From a processing point of view, power consumption is a huge consideration no matter where it is happening. But the dynamics are very different in a fully electric vehicle, which requires tradeoffs between energy consumption and performance. Off-the-shelf processors won’t work in EVs, or in the infrastructure equipment. Although it’s technically possible to use off-the-shelf hardware to run exactly the same algorithms in a piece of infrastructure equipment, such as a roadside sign, for object detection and classification the power considerations alone are not workable.

“If you’re at an intersection in New York City during rush hour, you’re detecting so many more objects, but the process is the same so the compute requirements are huge,” observed David Fritz, senior director for autonomous and ADAS at Siemens Digital Industries Software. “A lot of people would like all of that computational power to be driven from a solar panel, but we’re not there yet. That solar panel would have to be the size of a football field. So we’re back into the considerations on the vehicle side. We have to run AI inferencing algorithms. We need to have a lot of training, and there are many processes happening. All of that consumes a lot of power. How do we get the power consumption down? It means the idea of off-the-shelf hardware components is not going to fly, especially when they’re $20,000 to $30,000 a box, and there are 6,000 intersections in New York City.”

That clearly doesn’t work. “The way to design these things is to design them with power in mind,” said Fritz. “You need to be scalable but the solution needs to interact with, and be congruent with, the onboard solution that is designed for very low power. It’s about translating the low power emphasis in-vehicle into smart infrastructure.”

That can save enormous amounts of power, both in the vehicle and across the urban power grid.

“It’s important to remember that a lot of power consumption is from the electric motor, so being able to optimize traffic and traffic flows will make a difference,” said Kurt Shuler, vice president of marketing at Arteris IP. “In terms of vehicle infrastructure, having the big Waze brain be able to actually direct your car and optimize all of that would reduce power. It’s like fleet management for FedEx, but for everybody.”

To be sure, smart cars capable of communicating with one another and its environment open up the possibilities for safer roadways, more efficient travel, and better driving experiences. But it also will take time to roll out.

“V2I, and on a broader level V2X, requires the implementation of supporting intelligent infrastructure outside of the vehicle,” said Lars Ullrich, vice president of automotive at Infineon Americas. “Pilot programs already exist today, but we are still in the early stages of understanding the requirements and capabilities of existing technologies. That puts wide-scale adoption at least a half-decade or more in the future. Semiconductors play a significant role in the deployment of V2X when we consider the concept of the car as a connected network on wheels that requires sensing capabilities, connectivity, communications, and security. Everything must adhere to high-reliability automotive standards.”

Security
Getting this to work is only part of the challenge. An equally daunting task is figuring out how to make it all work securely.

“You have to make sure that the car that is talking to you is actually a car, and that it is a genuine car from a genuine manufacturer, not some hacker or adversarial entity trying to send fake signals or jam your reactions,” said Thierry Kouthon, technical product manager at Rambus Security. “This involves public key cryptography in order for cars to authenticate each other, so I know that the 100 signals that I’ve received from all those cars that are surrounding me are from actual cars from Volkswagen, Mercedes, General Motors, etc. In order to do that, you need two things. You need a root of trust to ensure that nobody can pervasively inspect your cryptographic processing, and your cryptographic operation while you’re processing them. You need that on both sides.”

In addition to the communication in the V2X context, there must be security on the inside of the vehicle. “Modern cars are often said to be more like computers on wheels than cars,” Kouthon noted. “This is because a typical car contains 50 to 100 computers for everything — steering, cruise control, entertainment, instrumentation, brake lights, you name it. It really has a lot of components, but 70% of those components have been provided by suppliers to the OEM, not by the OEM itself. How do you make sure that everybody that is supplying something to you is genuine, and that it’s not injecting the car with something malicious?”

The CAN bus, which long has been the in-vehicle communications network, typically does not include any security for cost reasons. “It’s not because the OEMs don’t know,” Kouthon said. “It’s because vehicle margins are slim. It’s a big bus, it’s a lot of wiring. To make it very secure would cost money and increase the car’s bill of materials, so the approach now is to put the security inside the electronic control units (ECUs), because each one of these ECUs has a computer inside the processor. Since we cannot secure the network between the processors, instead the processors are secured themselves so that when they talk to each other, regardless of the communication medium, they can authenticate each other.”

The data being communicated needs to be secured, as well. “You want to have the base properties in place,” said Jason Oberg, CTO of Tortuga Logic. “You want to encrypt everything, making sure that no one can steal the information. And then you want to make sure that all of your information is signed or is authentic. This means if I send something to another vehicle or to a building or to a sign or part of infrastructure, that it’s authentically coming from me. You don’t want an attacker on the side of the road, with their laptop with an antenna, injecting material into the system that causes it to think it’s me and causes my car to behave differently. Those are fundamental issues that need to be addressed.”

This may sound straightforward, but the problem is everything is so distributed that it creates a mess for key management. “How do you update the who is who,” asked Oberg. “How do I know that the keys that I have to say, ‘I’m Jason,’ are the most up-to-date keys? Who’s the central owner of all that?”

And because a vehicle is an actual device, these security features need to be stored at the hardware level. “Especially when you look at today’s infrastructure, someone could climb up a telephone pole and muck with something so it’s a different attack model,” he said. “So it’s really important to have everything routed down into the actual chip. That’s where a root of trust is really important, and those keys need to be stored securely somewhere. A root of trust provides a good framework for that, and updating them in those same environments is important as well. All the security complexities around how that’s all managed is important because you have to verify all of that to make sure you don’t make a mistake, make sure the system complexity isn’t overbearing, and this is going to be especially important for these V2V and V2X applications,” Oberg said.

Another aspect of this is that the level of communication that will be required to make it possible for vehicles to interface with any and all manner of infrastructure will be massive and 5G will be the conduit to making that happen, Jorg Grosse, product manager for functional safety, OneSpin Solutions, believes. “Artificial intelligence and edge-computing technology today will have to grow significantly while the level of verification needed will also increase. The complexities involved with making sure these communications happen safely and securely will be astounding. Formal verification will play a crucial role in the verification of the integrated circuits that will power this intelligent communication. The technology will go beyond simply making sure designs operate as intended but also must ensure that safety and security vulnerabilities and weaknesses are eliminated. I can see continuous product lifecycle verification being adopted as well, since the ages of devices communicating with one another will be a wide range. Not only will device age play a factor, but these devices will have to keep up with the constant evolution of security requirements.”

Regional rollouts will vary
Political infrastructure will play a role in how easily V2V and V2X are implemented, as well. The experience in the U.S. will be different from Europe, and both will be very different from China.

“A lot of that has to do with political systems,” Tu said. “China will probably be one of the leaders in that space. Their government is very active in building infrastructure and investing in infrastructure. Europe and Japan will probably follow, with the U.S. probably lagging a bit because we want the private sector to pick that up. We don’t look to the government necessarily to pay for those things, which is why the U.S. will follow. That being said, in the short term there are some innovative U.S. cities that are really trying to invite autonomous driving vehicles into their area with some early deployments.”

Not all features will be immediately available everywhere, even if the technology is ready. Tom Wong, director of marketing for design IP at Cadence, noted that because DSRC is already deployed, it is possible to buy a Volkswagen GTI in Germany today with built-in DSRC. So the SoCs and chipsets that support DSRC are ready, even if the external infrastructure is not. “The next step is 4G, and eventually 5G,” he said.

General Motors vehicles have offered 4G for several years, but it has had fairly limited usage. “They’re not using it to update software over-the-air,” Wong said. “They are using it for you to call somebody when you get locked out of the car, and they unlock the car for you — simple things like that. That’s clearly where the world is going, and there’s going to be geographical, country-specific 4G versus DSRC, which is highly controlled by the communications spectrum, and government regulation.”

For smart city implementation, there are still many issues to be worked out from the regulatory point of view, and that is important for the V2X trials to move forward, noted Ron DiGiuseppe, automotive IP segment manager at Synopsys. “Public trials supported by different localities/municipalities/cities will make sure the regulations can see what’s involved, and that the regulations match the industry deployments. While it is early days, trials are underway, and that’s a good thing.”

Again, this speaks to the tight coupling of V2X with the evolution to fully autonomous vehicles. “The common denominator in terms of technology is 5G wireless,” DiGiuseppe said. “For V2X, they want to use 5G for the communications. In fact, for the 3GPP implemented specific features in 5G that would make it more useful for automotive — one of which is improved reliability, and, of course, improved bandwidth — they’ve also added some prioritization. If someone in the back seat is streaming a movie for their rear-seat entertainment, the last thing you want is that streaming data to interfere with the safety-critical data. The V2X, or the autonomous driving 5G, has specific features to allow it to prioritize the data, and that helps for that safety application.”

Autonomous driving uses 5G in very high-resolution locations for route planning, and V2X uses 5G for V2X communications. But that bandwidth needs to be shared with the streaming movie, which makes this especially complicated. To confuse things further, with V2X trials happening in a growing number of cities globally, actual deployments will vary.

“There are a number of automakers that are already committed to V2X deployments into the telematics control units (TCUs),” DiGiuseppe said. “Then, the roadside units on the infrastructure side that communicate to the streetlights or the parking garage are implementing 5G radios. Those hardware implementations need to be completed and rolled out more broadly. It’s very clear that China has a little more freedom to define these smart cities, as opposed to the United States, which is more a case-by-case adoption on a community level. China has made some announcements about its smart city initiatives and has highlighted that it’s building a smart city from ground up. If you’re going to create a new smart city, then you have more flexibility to plan it.”

Overall, deployment of V2X infrastructure is contingent upon supportive regulations, major investments and development of critical technology for autonomous vehicles. “As we move toward higher autonomy levels in cars, more and more countries begin to engage in pilot projects that test the application of V2X in real life situations,” Infineon’s Ullrich said. “In the U.S., there is growing momentum toward deployment of V2X, especially as its benefits and ease-of-use become more apparent. There is some degree of freedom in the U.S. that enables early pilot applications, designed to understand and validate the technologies required to make V2X a reality. The country’s entrepreneurial nature has pushed regulatory bodies, start-ups and tech giants alike to work together in a tight-knit ecosystem. Many U.S. carmakers have committed to supporting V2X capabilities in new car models. The state of Michigan launched the Michigan Office of Future Mobility, which leads the strategic coordination of all mobility-related initiatives including infrastructure. Places like San Francisco, Atlanta, Colorado, Pittsburgh and many others also have rolled out pilot projects for autonomous vehicles and V2X.

Similarly, in Europe, V2X testing is taking place across many countries. There is widespread support from private companies and initiatives such as the European Union’s C-Roads project. 5G-ready tests also are taking place in anticipation of widespread rollout of 5G in the region.

Finally, China has numerous projects aimed at achieving mastering intelligent assisted driving. The country announced plans to install V2X technology on 90% of its highways.

Conclusion
Smart cities and smart cars are both making progress. The benefits of making this work are obvious even if the rollouts are uneven.

“C-V2X is one of those rare sensors in the vehicle that becomes more valuable with time, so the biggest challenge is to take that first step because more valuable means that the more cars that have it, the better it is,” said Qualcomm’s Misener. “It’s like herd immunity. If I can come to an intersection, and can say with some certitude that a whole bunch of cars are equipped, it gives me more safety. Ford has announced in the U.S., and a whole host of auto OEMs in China have announced, so that first step is happening. After that, people will start to see the benefits. Some of the benefits from small market penetration, but key infrastructure is what’s necessary to accelerate it. Once the safety benefits become known, and more OEMs happen, then in due time, when it’s a standard feature — or at least a feature that you want to buy into a car — then we’ll see even more benefits.”

Still, there’s a lot of work to be done to make that possible.