5G Connected Vehicles and Autonomous Driving

To learn more about Intel’s work with 5G, connected cars and autonomous driving, watch the webinar.


Good morning. Good afternoon, everybody. Thank you for attending today's webinar, titled 5G and End-to-End Enablement for Connected Vehicles and Autonomous Driving, presented today by Intel. My name is Iain Gillott, I'm the Founder and President of iGR, and I'll be moderating the discussion today as we get into the content.

We have three speakers today. And I'm actually going to-- we'll get them in a few seconds here-- I'm going to actually let them introduce themselves in a few seconds. But we have Caroline Chan from Intel, Geng Wu, also from Intel, who's an Intel fellow, and Patty Robb, who you can probably guess is from Intel. So we will-- I'll get to them in a second.

What I want to do now is start with a little bit of housekeeping. I'll walk you through the agenda. We'll introduce the speakers, and we'll get into the structure of today's content. Rather than have each presenter present, you know, 8 or 10 slides for 10 minutes or so and then get to questions at the end, what we're going to do today is try and cover four different areas of the 5G and connected vehicle, autonomous driving ecosystem, if you like.

And what we're going to do is we'll cover each topic, and then we'll pause for questions from you, the audience, at the end of that particular subject. If you have any questions, please use the Q&A tab on the bottom of the control-- of the panel there, [INAUDIBLE] 24 panel. You'll see there's a button down there that says Q&A. If you push that, then you can submit questions to us, and we'll take them as we go along, OK?

So what we're talking about today, our first topic will be, why is 5G critical to supporting and enabling fully-realized connected vehicles? The second topic we'll move onto will be, what are the technical requirements, and how do they differ between connected and autonomous vehicles? And there is a difference between those two. Again, we'll be taking questions after each of these sections.

And what is the role of 5G networks for CV and AD? How do edge computing and network slicing apply? And then finally, what role does Intel play in the development of 5G, connected vehicles, and autonomous driving? We'll have another Q&A section at the end. And we will move through that way.

So with that, just a couple more things. This is being streamed through your computer. There's no dial-in number. For the best audio quality, turn up the volume. You can find additional answers to some of the common technical issues located in the Help button at the bottom of the screen there. We are recording the webinar, and it will be available on-demand about 24 hours after the event.

And as I said, we will have questions. If you do have questions, just use that Q&A button down there. And with that, let's get started. So I want to introduce our speakers again. And I'm going to ask each of them-- we'll start with Caroline, then Geng, and then Patricia, Patty-- ask them what you do at Intel, and what your interest is in 5G, and what your interest is in connected vehicles and autonomous driving. Caroline?

Hi, everyone. This is Caroline Chan. I'm part of the Intel 5G team and manage the 5G infrastructure side. My interest in the 5G connected cars is primarily around how we're you using the characteristics of 5G and the fact that we have now going forward into virtualizing the network as well as providing the edge cloud, all this combination, I think, will make the connected car very interesting in the 5G era. I'm looking forward to this discussion today.

OK, great. Thank you. Geng?

Hi. My name is Geng Wu. I'm a part of the Intel for Wireless Next Generation Standards Organization. In particular, my focus is on wireless standards, global wireless standards development and ecosystem collaboration.

My focus, certainly, is now, I see industry shifting focus from enhanced mobile broadband to the internet of things and bringing billions of devices, including connected cars, to the network. My current focus is the 3GPP Standards development, 5GAA, and AECC, which stands for Autonomous Edge Computing Consortium, which are the basic and essential elements to support the industry.

Great, thank you. So you're the acronym guy. [LAUGHS]


Good morning. I'm Patty Robb, and I lead the Intel Silicon Valley Innovation Center and Worldwide Vehicle Lab. And my focus here is on autonomous driving and also leading our 100-vehicle fleet. And I'm really passionate about 5G because it brings connectivity to everyone around the world and also autonomous driving.

And really, these two together are important to me because of the positive social impact it will have for the world. And I was going to talk, you know, a little bit about this, but I think the interesting thing that people may not know is that from a big-picture perspective, Intel's really involved in a lot of things on 5G.

And for me, I think the other piece is, our Mobileye technology that we recently acquired is already in over 24 million vehicles for today's ADAS systems. So if you're thinking about a car out on the road today with driver-assist functions, such as automatic emergency braking, lane keeping support, traffic sign recognition, and more, you're likely to already be experiencing our Mobileye technology.

And studies have already shown that ADAS technologies are reducing fatal accidents by 30%. And for Intel, the Mobileye technology is foundational for autonomous vehicles. So I'm super excited to work in this space because of the positive and transformational impact it'll have on society.

OK, great. Thank you. OK. So with that, let's dive into our first area here. And we'll talk for a few minutes here. And again, I want to hear from each of the panelists. Why is 5G critical to supporting and enabling a fully-realized connected vehicle?

I mean, when you think of connected vehicles, obviously, you could be streaming something to the vehicle itself. So with autonomous driving, where does 5G fit into that? Why do we need the connectivity there? So, Caroline, I don't know if you want to start. I know you've got the 5G infrastructure view for Intel. So why 5G in autonomous vehicles, connected vehicles?

I think the first thing that happens with 5G, as Geng has alluded to, is that it is a standard that goes beyond the mobile broadband, traditionally, what the cell industry has been focusing on. It is taking great steps to address the vertical industry, such as autonomous driving or connected vehicles.

That's one aspect, so the low latency, the high throughput, and the ability to do network slicing really makes and transforms the network to be much more cloud-like than a purposely built network focusing on mobile broadband or voice communication like we have in the past. So that's one aspect.

The second aspect is that the whole momentum of Nav SDN has been sweeping through the network, starting from the core network, now it's going down to the radio access network. What it has started, essentially, it makes the network behaving much more flexible and software defined. So you could actually choose to add more of the smart, and the computer, and the storage towards the end.

So things that come with the connect car that before had been way too expensive or not technically feasible become a viable option for both the car manufacturers as well as the network operation-- like HD map download, like entertainment, like road hazard condition update, that Patty can share a little bit more. And just like Patty said, we've seen that the ADAS has significantly reduced the road casualties, the accidents, and so on.

So by working through the industry, like the consortia like 5GAA, like AECC, the Automotive Computer Consortium, we're starting to look at, if the operator today has benefited from Nav SDN, an overall more efficient network, how will we then relate that to the verticals, like the car manufacturers, especially combining the fact that 5G brings out the low latency, the high throughput, and also the ability to have alternative spectrum, not just in the sub-6 as well as above 6 gigahertz, which makes, to put it bluntly, the whole much broader access to cellular technology, to the spectrum, as we never had before.

So to me, at least, it becomes very interesting to Intel as a whole, we have all the portfolio that you just heard from my partners, like Geng, like Patty, to bring in the technology of Mobileye, the technology of AI and machine learning, and the cellular heritage that we have. That makes it a much more compelling story for connected cars. I think this is why we are so passionate and interested about exploring this together with the ecosystem.

OK, great. So--



I'm just going to [INAUDIBLE] topic, just one thing that people may not know is that autonomous vehicles are designed to work with no connectivity at all. It's a safety feature. And they need to be able to operate in all environments, right? So the vehicles need to be able to operate in an area that has no cellular coverage, like remote and rural areas. So really, connectivity enhances that experience.

Right. And actually, you're leading into the next question I was going to ask, is, you know, I live in Texas, I live in Austin. Caroline's up in Dallas. She and I don't have to drive very far west before we end up in the middle of nowhere with a lot of cows. But there are certainly some very big roads.

And if you drive from Austin to LA, for example, you spend most of your time on I-10. It's a big highway with a lot of vehicles on it, but there's certainly not a lot of infrastructure out there. So, you know, Patty, what you were talking about was the need that as we move out of coverage, 5G is not going to be everywhere initially, then certainly, the autonomous vehicle needs to be able to survive on its own, right?

It is designed that way. That's what I want to make sure everybody understands is that the autonomous vehicle doesn't require connectivity. And that's actually designed that way as a safety feature because it has to be able to operate in all environments.


So it's the design from the beginning.

Yup. So one last question on this subject, and then-- actually, we've got a couple of questions from the audience as well, which I'll get to in a second-- but Geng, I wanted to come back to you. And you're involved in a lot of these standards bodies, both on the 3GPP side and on the connected vehicles, autonomous driving, et cetera.

And the obvious question is, well, which comes first? Do you see 5G coming first or autonomous vehicles? Are they coexistent? You know, how are these things developing from the standards point of view?

They are certainly in development in parallel. And to be more specific, there's several important things about the 5G, about the wireless industry working together with the automotive industry and also supporting and guiding requirements.

So there are several [INAUDIBLE] specifics about the 5G. Number one is that it gives the combined industry this global scaling interoperability, which, in terms of economy of scale, and also the user friendliness for people to have a consistent experience when they travel to different places.

Another thing is that 5G is a re-do of wireless technology. The many benefits, many technology advancements can directly benefit the automotive industry. But fundamentally, as Patty pointed out, the complication is a supporting mechanism to augment the sentient capability, to go beyond just the car's immediate surrounding environment. However, ultimately, it's a supporting mechanism, and it's expanded sentience.

Another thing important to 5G and also to the automotive industry is car, as we move forward, becomes a platform. And many value-added services can be introduced there as connected cars. And that's where 5G standards development is currently focusing on.

OK, great. And I think I just put up a slide here with some of the standards bodies you guys are involved in, which we'll come to later on. So before we move on from the subject-- we'll get into some more detail, obviously, as we go through here-- a couple of questions.

I think, firstly, this one is a good one for you, Caroline. And again, for the audience, if you do have questions, use the Q&A tab. We can take them as we go along here. I'm just going to read this one out for you, Caroline. Can you speak in more detail to what it means when 5G arrives?

With 5G networks, what will be the impact to IT infrastructure, and what does it mean for intelligence-- sorry, integration between data centers, towers, and antennas? So more of a generic 5G question rather than specifically to CV and AD. So if you want to give a few seconds on, what can we expect when 5G arrives, probably later this year, early next year?

Yeah. Yeah, 5G solves world peace. No, I'm kidding.


The way that we look at it-- and this is actually a really good question. So since I take on this role, I've been going out to talk to different CIOs and enterprise IT departments. One thing that really comes up is that 5G shifts from the traditional cellular only. Like, it's not the IT world, and I don't have nothing to do with it, to a very heightened interest from the IT side to look at, can this become a connectivity play to my IT?

So the way that we used to look at this, is like-- Iain, you and I have been talking about this edge car concept for a long time. We would talk about connectivity that comes with the cloud. But for a lot of the IT departments, the CIOs, it really becomes, I have my IT infrastructure, and I need a reliable connectivity, a connectivity not just ever present, but connectivity also suited to my particular application needs.

We mentioned [INAUDIBLE] mentioned high broadband. To them, it's somewhat like speaking Greek to them. But if we can tell them or find a way to address their particular services and application, their IT needs, if it comes with a consistent connectivity with the right amount of SLA, some kind of guarantee, that's much more music to their ears.

So we started tuning the 5G messaging towards the IT side, knowing that, for 5G to become successful the way that everyone wants it to be, you really do need to start addressing some of the verticals, the enterprise requirements. What are the requirements?

They really look-- everything, to them, is a cloud. There's a lot of microservicing. You need to be able to spin up microservices in minutes and not days and weeks that typically runs today. You need to be automated. You need to be able to orchestrate different-- the network as smoothly as they do it today on a cloud space.

So we started working towards that with our ecosystem. You see a lot of the announcements coming out-- [INAUDIBLE] it's really to start saying, the network needs to be run more like the cloud. It needs to be run, the platform as a service and network as services, and needs to be able to manage itself and automate the entire microservice spin up. So I think the connected car that we wanted to achieve needs to be aligned with that vision.


It cannot be, right, the way the cellular does today. So anyway, so I think that this is probably one of the big problems that we had to solve.

OK. Great. Well, let's move on to our next section here. Let's get into some of those more technical requirements here. And I do want to talk about how the technical requirements differ between connected vehicles and autonomous driving, autonomous vehicles.

And this is probably a good point, good time to say that we've actually written a series of three white papers, which I'll give you the links to at the end, with Intel that address all aspects of the autonomous vehicle-- sorry-- autonomous vehicle, connected vehicle ecosystem, landscapes, and capabilities and also, really the role for 5G within this.

So I've got a quick slide here which really says, what is a connected vehicle? Well, it can have a 3G, 4G, or 5G cellular radio. Or it could have some sort of dedicated DSRC short-range communications that does vehicle to everything. So this is where the vehicle can talk to the stoplight and things like this.

And, of course, you can get the cellular and the connection in a vehicle today. A lot of cars have 4G Wi-Fi hotspots within the car, et cetera, et cetera. Automated vehicles, of course, integrate varying levels of autonomous driving capabilities.

And the Society of Automotive Engineers has assigned these different levels to different capabilities, basically, all the way up to level 5, well, from level 0 to level 5, where level 5 is full driving automation. So that's when you sit in the back seat, and the car drives you home. And I think we can talk about this in a few seconds here. So that's really the difference we're talking about between connected vehicle, autonomous vehicles. They are different here.

To give a little bit more detail, there are different standards. And this one we do want to get into in some more detail here. But there's C, cellular, V2X, which is part of Release 14, and then there's also, as I mentioned, DSRC, which is actually its own standard based on 802.11. It's actually been around quite some time, but it really doesn't seem to have moved that much.

So with that, let's open up the discussion again. And we do have some questions around this which I think will be good ones here. So the first one, actually, I'm going to take is a question here. It's actually from Joe Madden. What can you do with a 5G car that you cannot do with LTE or today's capabilities? So who wants to take that one?

I think we can talk about a number of use cases. I don't know if you've got any particular slide that we might be able to pull up that could allow us to show that. But basically, you'll see a whole flurry of use cases that can come in, anything from high-volume data upload from test fleets that can benefit from millimeter waves.

You can also see within there, let's say someone wanted to do a remote control of an EV with overload latency. If the vehicle needs maintenance or repair, you can do the remote control of that vehicle over 5G over low latency. Over-the-air updates for security patches, just basically pushing those to the vehicle much more quickly.

You'll also see a lot of things related to the entertainment of the passenger and increasing what the passengers are able to do over it. So for instance, one of the things in 5G is that the rise of the autonomous vehicle industry is really going to create one of the greatest expansions, Joe, of our consumer time available for entertainment that we've seen in a long time. And as passengers shift from being drivers to riders, their connected device time, including video viewing time, will increase.

In fact, we've actually done a study and there's a survey that shows that Americans spend more than 300 hours per year behind the wheel. And 5G's going to be an essential part of delivering new experiences to the customers that they're going to demand when they're riding in the vehicle.

And Intel actually predicts a new passenger economy will emerge to support the idle time when drivers become riders. So the economic opportunity from this is, it'll go from $800 billion to $7 trillion as autonomous vehicles become mainstream. So you think about it as mobility as a service will disrupt long-held patterns of things such as car ownership, how people get around, how things are maintained, and operations and usage, just to name a few.

OK. Yeah, so what you're saying is, really, the fundamental difference is, the main aspects of 5G that we benefit from are increased bandwidth and low latency. And those play directly into the autonomous vehicle, connected vehicle landscape.

So today, you may be able to get a couple of megabits to the kids in the back seat to watch something on an iPad. But you are limited because of that higher latency and, frankly, limited bandwidth. The use cases you talked about there, Patty, have all been using that higher bandwidth and low latency to the benefits, right?

Yeah. May I, actually, add--

Yeah, totally. Yeah.

--one comment. So from a standards development point of view, 5G is not a standalone event or activity. It's actually both about harnessing the capabilities we have, expanding them, and also about adding new capabilities.

To be more specific, developing the features and platforms that are more attuned to support the machine type of communications, a connected car is where the human communication and the machine communications are coming together. So it's about how to expand the machine's capability when things are connected, they can do-- in fact, they actually benefit and multiply-- and communication with communication computing getting together.

Another thing is that for the connected car and autonomous driving, we basically need a new type of interface, the network architecture, and also cross-industry collaboration as to where 5G fits sparsely into the landing zone. And there is a whole set of requirements that Patty mentioned, the 5G as a continuation of the 4G development, and a lot of things still need to happen.

Another aspect is that as the capabilities of the future connected cars expand, there's going to be a need for more spectrum. And 5G broadens the spectrum usage to beyond the typical cellular frequency band, and that's where the industry can grow further and the consumers can benefit from the expanded capabilities.

OK. So Geng, you raised a really good point just then, and we've actually had a question. And I'll read the question for you, and I think you can just continue that thought process and answer it here. But what are those new bands that are being used in 5G other than the 5.9 gig? What is the data throughput that is required for connected vehicles? So do you see other bands specifically that are being discussed right now within the 5G and connected vehicle, autonomous vehicle discussions?

Well, certainly the ITS band, the intelligence transport system band at 5.9 gigahertz, is the one for the initial technology development and trials that are widely used, supported by the cellular V2X technology for safety and also the [INAUDIBLE] for the safety messages for the cars to send to other cars.

And moving beyond, the use cases, expending the broadcasting certainly is no longer sufficient. There's going to be the unicast capability. And so in terms of a frequency band, the sub-10 gigahertz band is one area of high interest. And there's multiple possibilities.

And also, there's, for example, a 60 gigahertz band, about 60 gigahertz band, where the millimeter wave transmission can provide massive capabilities.

OK, great. So there's a second part to that question, Patty, which you started to mention a little bit, and I think probably you'd be a good person to answer this one. Is what sort of bandwidth do we require for a connected vehicle? What are you anticipating in the future? I mean, obviously, low latency we know, you know, sub-5 milliseconds. But from Intel's point of view, what does broadband to a vehicle actually mean in terms of megabits per second, 1 gig, whatever?

I think I'll let Caroline expand on that as well. But I think from our side, Intel's playing an end-to-end role in the development of 5G and connected vehicles, right? So we've got our end-to-end play. We're focused on 4G as well as 5G New Radio and the car and the network. We've got our edge computing via MEC with a mobile edge compute server.

Our ecosystem partners, we're working with them to develop V2X applications for traffic management. We're participating in all of the standards bodies. We also have live labs with trials going on around the world to check out the different bandwidths and use cases for smart cities and smart parking.

And we have our Intel infotainment platforms, providing the best-in-class user experience. And we're also leading cloud and data center technology, provides best-in-class compute, storage, and network capabilities. So we're really looking at this as an end-to-end solutions play. And I'll let either Geng or Caroline add in. I know we're talking down to a millisecond of latency.


And they may want to add some more to that.

So Caroline, give me a bandwidth number.

So the bandwidth number, I think it really varies between what type of application that you need to run. So it's probably a range rather than a solid number. But one thing, though, I do want to bring to focus is that what we talk about here, the cloud is not the traditional cloud that you're thinking of. We're moving that much closer to the edge.

So things like, [INAUDIBLE] that we've been partnering with Crown Castle, Vapor IO, start building some of the mini data centers, edge data centers, up in the tower. It is to get closer to where the cars are, where the users are. In that case, you are looking at being able to take advantage maybe a short-range type of a communication by the higher bandwidth, much, much broader spectrum that Geng was talking about.

So I would say, don't look at it as a traditional cellular case, but look at it as a different way to address a problem, whether you'd want to download an HD map, or just infotainment download, or updating some of the computer vision images that you're getting.

I think the question that needs to be looked at is, how do we cost-effectively deliver this data to where the cars need it? Patty said it, right? This is not about autonomous driving. This is really about assisting the cars to make the decisions, assisting the users, like HD map download. So we really are having a different way to solve this issue. And the edge compute becomes a very important part of the equation.



And that is a great lead-in-- oh, yeah. Yes, Geng.

Yeah, let me add a few numbers here.


Basically, as the use cases evolve under trial, the automobile industry gaining experience with trials that we're participating, the data rate, the demand for bandwidth, and also, to be more specific, for data rate increases.

And for that very reason, the standard for the moment, actually, currently in 3GPP is we're doing it in three phases. For the phase 1, Release 14, we are targeting 1 to 10 megabits per second. For Release 15, the target increased to 10 to 50 milliseconds. And in phase 3, which is falling into the 5G NR and the future releases, we basically are targeting 50 to 500 megabits per second.

OK, great. Good. Good, you answered it. OK. So all this discussion is great. And let's lead into our next discussion. Let's get a little bit more on the technical side here. Caroline, you mentioned edge computing. I think Patty, you mentioned new architectures. Somebody at the beginning mentioned network slicing. We want to get into all those types of things.

The other thing I have to say is we've done several webinars with this format. I don't think we've ever seen quite so many questions-- [CHUCKLES] coming in on the subject. This is great. And as an audience, you're doing really well. In fact, we've got more than enough here to keep that discussion going. So I'm going to try and work in these questions as we get into the more technical subjects here. But if you do have questions, please feel free. We've got some good ones here.

So I'm going to put in a slide here, network edge. And Caroline, you'd mentioned edge computing, and somebody mentioned getting the content closer to the vehicles, et cetera, et cetera. So do you want to expand on that a little bit? And specifically, let's talk about what Intel here is doing with edge compute because you and I've been talking about this for the last two or three years now, right?

Yeah, we have. So we started looking at this as saying that, to change the economic equation of building the cellular network, being able to monetize it, we need to introduce a behavior that has benefited the cloud guys in the past. So we've been working with the ecosystem. And Iain has been a great partner of ours, developing different use cases, developing, more importantly, our wire model.

We started taking on different verticals. We now have a software development kit, SDK, that's available on our website that people can come in and download through a simple click-through process, which can go with a select set of hardware platforms that you can acquire. These are simple, off-the-shelf platforms. What we want to do is to start developing different use case models that can go into a fast proof of concept and deployment.

In the process, I mentioned we work with Crown Castle, Vapor IO, we work with other companies that really start looking at how to solve the issues of edge computing. It's not simply by bringing the smartness to the edge. You actually need to solve some issues that you're very aware of, like security, like security in the mobility.

As you're moving the data from point A to point B, how do you quickly transfer the security certificate from point A to point B? A lot of these are issues that really need to be solved before this becomes widely available. And we're working on that today. We have some successful proof of concepts and trials, and we've seen some early deployment going now coming up.

If I may do a quick advertisement, in the Winter Olympics, we'll be demonstrating some of that in Pyeongchang. If you happen to be in the area, please come by. I'll be there, and I can show you some of the things we're working with the operator as well as with the car service provider to jointly bring this forward into a proof of concept and a trial that shows they it can be done. And some of the learnings we have published in the white papers.

Was that you, Caroline, I saw practicing the ski jump the other day on the video?

Yeah. They call me the eagle. Yes.


So I want to follow on with this, exactly the point you're making here. And we've got a question here, which is good. Where is computing or applications most likely to happen on the edge, and which will happen in the cloud or a centralized data center? So, and let's keep this to connected vehicle, autonomous vehicle. But which applications do you see being pushed down to the edge, and which are going to stay up in the cloud? Patty, you want to take that one?

I think there's going to be a number of different things. You know, if you-- I'll put the picture back up here, and I'm going to invite others to join me as well-- but I think there's a number of different applications that you're going to see over time as we look at the use cases for 5G. If you look at the picture, you see the entire network involved that I talked about. You've got the mobile edge compute server. You've got the cars, different things like this.

And one particular application you might have is around smart parking, right, and always having cameras looking at the parking spots to know, you know, at the edge, if there's parking spots that are available, and then sending notifications to users that there are spots available.

And obviously, at different points in time, autonomous vehicles might be out driving around, and at different points in time, they'll be parked. So I think having that automatic communication and rapport with the systems is something that's important as vehicles dock, to go ahead and take on maintenance or other things like this.


I think other--


So I'll let somebody else add to this as well.

Yeah, typically, the applications that require the real-time responses on the interactions, interactive, the process has to be to the edge of the network to actually achieve the required latency. And there's also tremendous value for the bigger cloud to provide more background information.

And those parts, for example, your location, the environment, and the facilities around you, et cetera, et cetera, and that's part [INAUDIBLE] greater geographic area, and that part can be still residing in the cloud, in the deep cloud.

Right, good. And I know that from the discussions I've been involved in with the edge computing folks is that that's a big discussion between the architectures is, how do you decide? There are applications that will reside at the edge. There are applications that will reside in the cloud. There are hybrid applications that will be in both. And frankly, there's applications that will start at the edge, move up to the cloud, and move back to the edge, although it may be a different edge.

So I know that's a lot of work that's been happening there. OK, so I want to move on to one other subject here. And I want to put up a different slide here and work in some more questions. But one thing I do want to talk about, we hear a lot about smart cities.

We hear smart cities, and we hear connected vehicles, and autonomous driving, intelligent transportation. Sometimes we hear them in the same breath. How do the needs of smart cities change the equation, or do they complement the needs of autonomous driving, connected vehicle? Anyone want to take that one?

I think, Patty. Patty, feel free to jump in. What I did when I was in China late last year, I attended a conference where they were making an announcement that one of the cloud service providers, Alibaba, in their headquarters was able to use the smart city application to essentially reduce the traffic congestion in the peak hour by 13 minutes.

It sounds like it's small, but it's actually tremendous, considering average traffic congestion in China, it's demoralizing to people. They were able to reduce it by 13 minutes on the average, and they are obviously working to reduce it more. So there is definitely, I think the smart city aspect, in terms of changing the driving patterns, the traffic congestion, the traffic light patterns, goes a great deal to solving the day-to-day issues that we all are facing, especially in some of the highly congested cities.


I think, Caroline, you brought up some great examples. I would add to that, I think they're complimentary, right? I think cities are looking at their smart city infrastructure. What network do they need to have in place? What do they need to do with their applications?

What do they need to do in place to be able to handle the large, growing number of residents that you'll see move into cities over the next few years? And how do they handle parking for autonomous vehicles? How do they handle more smart and connected services for their residents?

So I sort of see, even though the deployment of 5G and autonomous vehicles is independent and they aren't required from one another, I think cities will look to learn about and study both of these topics and look to find synergies between them so they can enhance the city infrastructure to be even better to serve its citizens.

Right. And actually--


--you bring up a good point. Yes. Yes, go. Go, go.

Yes. So smart city really is about a smart environment around us. And to make that happen, the machines, the people, and also the buildings, the infrastructures around us need to be smart. That's exactly the reason why the industries are working together on the V2X.

So it's not just one thing, but an x number of aspects. And also, it calls for the 5G infrastructure. And dealing with the smart city, you'll certainly need a new type of architecture, where the computing are moving you much closer to where the action happens.

Right. And actually, you started to answer one of the other questions we have, it's, what involvement do you see from governments with autonomous driving, smart city, et cetera? Importantly, how do you see this evolving? So do you see that government will get more involved in this? Is it going depend on where we are in the world? Or are they taking more of a hands-off approach?

Geng, do you want to take that one?

I think actually, globally, if you look at US, you look at China, Europe, there's a great level of support from the government, from the regulators. To be more specific, thanks to their support, there's many on-the-road trial that the industry, the automotive industry, and the telecommunication industry can work together.

And it could not have happened without the government and the regulators' support. And we actually would like to see more actions from the regulators as the city becomes smarter, cars become smarter, people become smarter. And then, the regulations need to evolve.

Right. OK, good. OK. So we've got a lot of questions here. They're still coming in, so I'm going to keep trying to work these in here. I think we've talked about the edge and some of those things enough now. Let's move now, let's talk more about what Intel plays in the development of 5G, connected vehicles, and autonomous driving. I know we've touched on this. And I'm going to show a slide with those use cases in a second.

But I wanted to try and bring in some of the questions now. As I said, we do have quite a lot here. And then, we'll try and-- you can maybe talk about some of the things you're doing specifically. So one thing that's come up again, and a couple of people have asked about this, is spectrum. We mentioned spectrum earlier. We talked about the 5.9. We talked about millimeter wave. We talked about some of the 10-gig bands.

So how are we going to decide which resource or which spectrum is going to be needed for sharing? How often is that going to happen? Obviously, it's a dynamic thing. But how do you see this happening between the needs of the end user, the connected vehicle, the smart city, the IoT, et cetera? Caroline, do you want to take that one of what you're doing there?

We haven't-- so what we're doing today, obviously, not being in the regulatory business, is to make sure that whatever spectrum becomes available, we are able to participate. CBRS, we are working on the CBRS. In conjunction with Geng and the rest of the Intel team, we developed a prototype in order to facilitate the spectrum sharing aspect. We work on the millimeter wave. We work on the sub-6 gigahertz.

I don't know what the answer is. I'm not in that regulatory side of the business. What we do know is that we wanted to ensure that the network infrastructure technology that we're developing in conjunction with our customers can meet the ultimate, whether it's sub-6 or millimeter wave side of the business. And [INAUDIBLE] CBRS is definitely very intriguing to us. I don't know, Geng, if you want to add more comments to that.

Yeah. So we need to actually keep in mind that many of these connected car use cases actually can and are covered by the LTE and the 5G interface capabilities. So many of the links, many of the communication requirements will be met by the existing allocation. And by the way, in the 5G era, we basically, one thing very positive is the governments around the world, actually, are allocating a massive amount of spectrum for future mobile industry growth.

Another thing to keep in mind is that for the specific ITS, like, the specific band allocated to mobile, we certainly would like to see global allocation for standardized allocations so that the machines and the infrastructures can interoperate and also have the economy of scale, and down to specific target band.

There's many business issues that still need to be addressed. For example, who owns the infrastructure? Who runs the network? And that's a part, as we move forward in the trials and the experiments, the industries together will learn more. And it is still in development process.

OK, great. So we've got a few minutes left here, and I've got three areas I want to talk about. And the first one, we've got multiple questions on this. We talked about cellular V2X, C-V2X. We also mentioned DSRC, and I had some comments on it earlier on. And how do you see those two either interacting, or do you see the dominance of one over the other?

As I think I said earlier, and somebody said in one of the questions, DSRC has been around for quite some time and really doesn't seem to have moved that much, whereas we seem to be talking a lot about, hearing about 5G and cellular connectivity. So how do you see the two between C-V2X and DSRC? What do you see is the role there? Geng, is that a good one for you?

Yeah, OK. So one. There are several key considerations in this. Number one is that when we talk about future connected cars, autonomous vehicles, it's not only about communication. In fact, it's about really the future autonomous machines and also the infrastructures that enable it, and supporting those machines, and how they actually augment the life experience or everyday, ordinary consumers. So it's not only about communication.

And another thing to consider or to keep in mind is that we have very diverse if you look at the use cases. The 5GAA has six key, leading use cases, and the 3GPP has four. But they all point in order to the one thing, which is the connotation need is very diverse, and we need a different type of communication with scalable capabilities, in terms of latency, reliability, and the data rate.

So another thing, the final thing here to consider is the requirements evolve, and the industry, actually, is still at the technology development phase. The requirements evolve and so do standards. Whatever we adopt needs to have a longer range of road map to evolve the capabilities of the interface and also the infrastructure to meet the car industry's requirement.

That's the reason why the industry, and from an Intel point of view, we are actually really, basically, focusing more on the cellular-based communication technologies.

OK. All right, good. So we've got a few minutes left here, and I've got a few questions. One that's come up a couple of times, so I think this can be a fairly quick answer, is a lot of discussion of, is there a link between autonomous driving-- and Patty, I think this may be a good one for you-- is there a link between autonomous driving, autonomous control of vehicles, and autonomous flight, control of drones, et cetera? It would seem that you've got a lot of the same capabilities there, except one takes off, and one stays on the ground. [INAUDIBLE].

Well, I think one thing you could see-- and this is something that's possible-- enabled through technology is you could see someone in a use case where you have an autonomous vehicle, and then they may have a drone, right, that goes the last mile with delivering a good or service to someone. So that's a use case that could be enabled through 5G.

You've got the autonomous vehicle, which operates on its own without connectivity. And then, you've got the autonomous drone. And obviously, connectivity enhances the experience in the autonomous vehicle. But then, you have the autonomous drone, which could work with a car or a delivery truck and actually take that good or service and take it the last mile to an end user. So that might be one use case that you could see.

OK, good. And yeah, I mean, if you imagine a drone flying through a natural disaster, a wildfire or something, to drop medical supplies, there may not be connectivity. So it would be truly autonomous at that point, just as the vehicle would be, off network.

OK. There's one quick question here, and we haven't raised it here. And Geng, I'm going to let you have this one. And then, we're going to talk about business models. But security, security of 5G networks. And obviously, if we've got autonomous vehicles running around, there's the potential for somebody to hack in there and cause mischief, injury, death. So I know you were involved in the standards bodies on here. But how's the industry-- and just maybe take a minute or so-- how is the industry addressing the security aspects here?

This is the paramount challenge to the technology, to the communication industry, to the technology community. So there's a safety aspect. There's the security, which is primarily the how to prevent the unauthorized transmission and, for example, just imagine someone put a jammer there, and they're broadcasting, they authorize the message, and the whole highway shut down.

So those are the real concerns and considerations. The industry certainly is looking at every multiple liable for mechanisms. There's the interface itself needs to be more secure. And the application layer needs to add-- basically, it ultimately needs to safeguard the system, the integrity of the system.

So it's not going to be one single answer to this challenge, to this paramount challenge. And to be more specific, the wire distribution and also the high mobility of the moving [INAUDIBLE] that's actually added more complexity to the solutions. So this is where the industry is working on basically across multiple layers.

OK, great. OK, so I want to come to the last area here. We've got a few minutes left, so I need to start wrapping up. We've had a lot of questions on business models. Where is the revenue going to come from? Who pays? And I'm going to assume that there are multiple sources of revenue potentially. So as a last wrap up-- and Caroline, I'm going to start with you, just to give you a little heads up-- is--



Iain, I apologize. Just one quick last point if I could on that last topic is--


--Intel's working regularly to make sure we're designing security into every layer, right in our silicon and our software and everywhere that we're working. And we're also working regularly across the industry and different standards bodies. And even in the automotive area, we're working with leaders in the industry to join us in a collaborative effort to continue to develop additional safety standards that we can all align to.

And we ultimately work on designing securely into the products. And we also are pioneering in the safety area with our RSS model. So we have a number of things that we're doing to basically make sure that we're launching a product that is secure and safe.

Right. Great. So let's wrap up here. Caroline, I'm going to come back to you. Business models. Where do you see-- and let's go Caroline, Patty, Geng, a few seconds each, literally-- where do you see the revenue coming from? What are the revenue sources in connected vehicle, autonomous driving? What's going to drive that revenue side of that cost of that business model? Caroline?

One thing that we're focusing on is that when you have a connected car, that means there's a lot of data. The data gets accumulated. The data monetization is a source of revenue that we can all look at. The entire ecosystem should be able to participate in monetizing the data. Like insurance state information, car maintenance information that ultimately benefits the repair shops and the car manufacturers. So monetizing data would be my number one vote.

OK, great. Patty, revenue. Where do you see the revenue coming from? And I know you've got the-- Intel's got those 100, level-4 cars running around in trial right now. So what do you see from that?

Well, I'll talk quickly on two quick things. Right, we do have-- thanks for bringing that up-- we do have our fleet that we're working with in our garage labs. And we've got vehicles equipped with our Intel and Mobileye-based computing systems.

And we're actually working on this fleet. And it enhances our ability to bring solutions to our customers even more rapidly than ever. And we can merge the data and the learnings derived from that test fleet with that of our co-developed fleets that we do along with our OEMs.

So I think we learn a lot about this, and actually, it makes our products better. And I think what I would say, in terms of some of the applications and services, I think it can go any number of different ways. I talked about that $7 trillion passenger economy, right?

And I think you could have things anywhere from applications that enhance the entertainment experience in the car, you could have restaurants that drive around with people in the car. I think you're going to see just an explosion of different types of passenger experiences that are created as well as new jobs that are created as part of this passenger economy.

Great. OK, Geng, you get the last word. Your top revenue source for connected vehicle, autonomous driving.

Ultimately, it comes down to the ownership and the operation of the network of the platform, the car platform, the infrastructure, and also the data. So basically, from a standards point of view, the industry has been working together to enable the basic capability to meet the possibilities, the different kind of business relationship and ownership of the networks.

Great, great. Well, I have to thank the three of you, Caroline, Patty, and Geng. We've got more questions. Even at the end, more were coming in. We could literally go on for another 30 minutes. But unfortunately, we do have to wrap up now. So I want to thank each of you for your time today. And of course, I want to thank Intel for sponsoring today's webinar with FierceMarkets.

You've had some great questions. We didn't get to all of them. So they have been recorded, and I'll take the time in the next few days to send emails out and then forward back to Intel as well to get those addressed. This webinar has been recorded.

You'll be able to access the recording within 24 hours on the same page you used to register for the event. I want to thank you again for joining. It's been a really dynamic session. And the level of questions, the input here has been terrific. Obviously, a very big subject. And look forward to seeing you at future events. Thank you very much.

Thank you.