Autonomous freight trucking coming soon

Markham interviews Dr. James Jeffs, the principal technology analyst for UK-based IDTechX.

What is Energi Talks?

Journalist Markham Hislop interviews leading energy experts from around the world about the energy transition and climate change.

Markham: 00:04

Welcome to episode 366 of the Energy Talks podcast. I'm energy journalist, Markham Hisler. Robotaxis get all the public attention as illustrated by Tesla's announcement last week that it will soon have a commercial model ready to go. But quietly behind the scenes, autonomous trucks are already beginning to take off. They promise to lower costs and improve safety for an industry plagued with driver shortages.

Markham: 00:30

As many as 500,000 level 4 autonomous trucks could be on the road in the very near future. I'm joined by doctor James Jeffs, the principal technology analyst for UK based ID TechX to discuss his recent report about the state of the autonomous trucking sector. So welcome to the interview, James.

James: 00:50

Hi, Malcolm. Thank you for having me back. And, yeah, looking forward to it.

Markham: 00:55

Yeah. This is this is very interesting. They, the light duty, whether it's cars or trucks or whatever, the light duty sector gets a lot of the attention. And why wouldn't it? Right?

Markham: 01:05

I mean, most people, you know, they have a vehicle that's that's what they're most familiar with and most interested in. But there are a lot of trucks on the road and different kinds of trucks and the all sorts of problems in the trucking industry. I mentioned some of them like high costs and and driver shortages. Maybe we could just start with an overview of what an autonomous what kind of autonomous trucks are we talking about?

James: 01:31

Yeah. Sure. So the levels that we use to describe the cars, so the Society of Automotive Engineers, they're 1 level 0 to 5 also applies for trucks. So when we're talking about autonomous trucks, most activity and the end goal is around level 4. So we're trying to get, driver out unsupervised trucks that can go from point a to to point b.

James: 02:00

That's kind of where the the top end of the research is, but there's also a lot of value around level 2 plus. So that would be trucks that can do supervised autonomy. So you can take your hands off, but you have to watch what the truck is doing and monitor it and supervise it. And also level 3. So perhaps, the truck can do fully autonomous driving on, highways, but it can't do the last couple of miles into a city or or whatever.

James: 02:30

So you can have the truck unsupervised on the highway with a operator chilling out in the back, maybe having a sleep if it's overnight, and then they can resume when the truck gets to a an operational design domain it can't cope with. So I think that's when we talk about autonomous trucks. Those are the the three kind of technologies we're looking at at the moment.

Markham: 02:50

And where are we at in the development of these technologies and their deployment? I I I don't see them. Maybe I wouldn't even notice them. If if I did see them, wouldn't know what I was looking at. But I my impression is we're not seeing a lot of them on the road yet.

Markham: 03:07

So where what's the state of the the industry development?

James: 03:10

Yeah. So all 3 of those technologies that I mentioned are in development. They're not deployed. As far as I'm aware, there's nowhere, where you can operate a truck with your hands off the wheel. Like trucks aren't being sold like that.

James: 03:24

You can't operate trucks like that. Whereas in the the car industry, in the US and in Europe, now you you can take your hands off the wheel as long as you keep your eyes on the road for certain cars that have been certified to do so. So, yeah, all these technologies are still developments, not not fully deployed. I definitely level 2 plus is closest because it's a very small step from what's already deployed, which is, you know, trucks already have adaptive cruise control and then keep assist and the kind of, ADAS functionality we're familiar with in cars. So to go from that to hands off, but the driver is still watching what's going on and supervising isn't such a big step.

James: 04:05

But there's a lot of work going on around level 3 and level 4 and a lot of testing with drivers in. So the the most advanced tests that we see are driver in with the vehicle operating autonomously. So the driver is supervising from, like, you know, the the they're from the development company. They're, part of the staff there watching it, making sure it's doing everything that it's supposed to be doing. They're they're specialists, but it's got a paid cargo in the back.

James: 04:35

So that's becoming more common in the US and China. And then the most advanced things we've seen are a couple of tests with, nobody onboard. So the vehicle is monitored remotely at data centers. It's not carrying a, commercial load in the back, but it's doing a, so I think the longest journey we've seen is is maybe around a 100 miles, 100 mile journey, autonomously. No one no one in the vehicle.

Markham: 05:01

How long before we get to full level 4 and these vehicles don't have a a driver in them, and, they're, you know, they're basically driving from one side of the the country to the other all by themselves?

James: 05:19

So the first part of that question is not too far away. When do we get to full level 4? That's not too far. The tricky bit is one side of the country to the other. So what we're seeing is a lot of testing on specific routes and the commercialization of of routes like Dallas to Houston or, Austin to, I don't know, Phoenix, something like that.

James: 05:44

And there's a a lot of value there in the you you pick these high volume routes that are between cities with a decent amount of mileage in between them, that isn't too busy. It's not like operating in the interstates like around LA. It's a a an achievable operational design domain. Those kind of routes we think will be commercialized very quickly. Like, within the next couple of years, we'll start seeing, driver's trucks operating commercial loads between the outskirts of a city, but, you know, one city to another around that Sun Belt region of the around those southern states.

James: 06:22

But,

Markham: 06:24

it seems to me, that the technology on these trucks is similar to what's on the, you know, the autonomous cars, the robo taxis, that are starting to appear, at least in the United States, Austin, Phoenix, are a few of the San Francisco, I think, all also has some of them. And there's a big debate over the what kind of technology. Is it LIDAR? Is it cameras only, like in the Teslas? Are there similar kinds of debates going on with the autonomous trucking technology?

James: 07:00

I I don't think so at the moment because I think everybody is more safety concerned. So the debate in the autonomous car industry is essentially around what's needed because these cars need to be affordable. If we put a lidar on and it puts the sticker price up by another $5,000, do we really, really need that lidar? Whereas the operating envelope for a truck is much higher. You're already working on a vehicle that costs into the 6 figures.

James: 07:31

And especially if you start talking about electric trucks as well, then the, additional cost of the sensor suite becomes quite small. So I don't think anyone's worried about trying to minimize the sensor suite and, you know, nickel and dime that yet trying to save everything they can. So we see we do tend to see most people still using LiDAR. I don't think I've seen anyone who's particularly a big advocate for the growing camera only for, trucks. But the we also see that the requirement shift a little bit for trucks, particularly the range, that the sensors can measure out to, and that's that's just around the stopping distance of the trucks.

James: 08:09

So you you see, like cameras that can have ranges of of a mile to a mile and a half, radars with range of, 800 meters, whereas on a vehicle, you're probably looking at more like 2 to 300 meters. Lidar's trickier to get out to those ranges though, without kind of breaking, like, eye safety regulations and and having the lasers too powerful to be used.

Markham: 08:35

What I've often wondered why the industry, doesn't geofence, roots. Particularly, you mentioned high volume roots, in the US. Would it make sense? Would it make the system easier to work if they were geofenced than there was, you know, data available every 100 feet or whatever the truck, required, so that it always knew where it was.

James: 09:03

So as in have the highways, like, what's that? Have have digital infrastructure on the highways to be monitoring?

Markham: 09:12

Yes. Exactly.

James: 09:13

I have my idea. I don't think it's necessary, and it becomes quite expensive if you want to start installing additional infrastructure on highways. And also a question of who's responsible for the upkeep and who's gonna pay that in the in the first case. And you do see certain departments of transport being more open to investing in highway infrastructure and and doing things like connected vehicles. So if you look at Georgia's DOT, they're very keen on trialing like vehicle to v two x technologies and, you know, vehicles communicating with different digital infrastructure systems.

James: 09:55

But you we haven't seen it really picking up much beyond the occasional proof of concept. So I I don't think it's something that that we'll see. I I don't think it will be needed for at first. I I think the having digitized infrastructure and having, vehicles for everything communication is something that we'll look at more in in 20 years time to finish off the the journey of autonomy and fill in the the last few blanks. I think there's a lot of progress to be made with just what we can do on the vehicle and the vehicle communicating to its own clouds at the moment.

Markham: 10:30

So these trucks have some fairly sophisticated systems, and they rely on computer chips. And there's been a lot of talk about the competition between the US and China and the US restricting the high end chips being exported to to China. Does that and be given the fact that it appears that in the autonomous truck sector, the competition is primarily between the US and China, does that that trade, war around chips, factor into this at all?

James: 11:05

I think that's definitely just a technology to war going on in in the chip sector for the the whole autonomous vehicles industry, not not just, not just trucking. But I think, you know, from the US's perspective, NVIDIA looks like the leader in terms of, what its top end chips can do and getting ultimate performance. And even in China, when you look at, flagship autonomous vehicles, they tend to be using NVIDIA chips. So from that perspective, it's it's good for the US. You know, and then got people like Horizon Robotics and Black Sesame, growing.

James: 11:45

But I I I don't think it's gonna be such a problem for the US just because Nvidia is dominant at the moment and Qualcomm entering the the arena as well.

Markham: 11:54

What about, the transition over to electric trucks? I take it that right now, the, most of the testing is done, with internal combustion engine trucks, diesel trucks. And, are there any advantages in the if the, you know, carriers begin to switch to electric?

James: 12:16

There's not an advantage in the, you know, any more than it is additive that you get the advantages of the electric drive train and you get the advantages of autonomy. They don't combine to give anything more than that. And it it creates additional challenges. So, obviously, there's the the range limitation with an electric electric truck. And if it if it needs to refuel or recharge, then and there's nobody on board, you that's a a problem to be solved.

James: 12:52

I mean, we that's partially solvable if you can do wireless charging like Elon announced last week with, WeRobot that they want to do wireless charging for their cyber gap, which makes a lot of sense. Additionally, in some ways, it can hold back the value of the autonomy. So one of the the big use cases or a really valuable use case later on is if you if you can do coast to coast in the US, a journey that's like 3000 miles that takes, 2 or 3 days by human driving. If you if you need to do that express with human driving, then you need 2 drivers on board. If you need to if you do it just with 1 driver, then they can only drive 10, 11 hours a day.

James: 13:37

It takes a long time. So if you do that route autonomously, but then you have to spend an hour every 5 to 6 hours charging. It it does take away some of the value of that route.

Markham: 13:51

And, what about the, the policy frameworks that are required for this? I I understand this has been an ongoing issue, ever since the autonomous, autonomous driving sector first emerged. How are the the various jurisdictions, doing in terms of resolving and, you know, redesigning its regulations?

James: 14:13

Yes. So we think that everything that is happening is essentially happening under exemptions, and the company is going to, you know, Texas local government and saying we want to do this. Can we have permission to do this between these two cities? That's pretty much how it's working for robo taxis at the minute. California is a bit more proactive than it has.

James: 14:36

You know, it's it's gone through a few different robo tech. It was gone through Cruise and Waymo getting licensed to do commercial deployments. But again, it's city by city, and I think it takes a lot of negotiation. I don't think there's anything yet at a federal level that is going to help roll this out nationwide. Same in China, the the agreements and the deployments are all city led.

James: 15:03

Companies will go to a city and negotiate what part of the city they can operate in. Obviously, in China, we don't know what happens behind the curtain as much, and we could have an announcement tomorrow that there's a a big shift and suddenly you can do a lot more widespread over China. And then in Europe, there's regulation work going on around consumer vehicles. So, UNECE puts out, like, the high level what you can and can't do on the roads have regulations in place for level 3 deployments, which is how the Mercedes s class got to level 3 in Germany. They also have, recently in September, they released a new regulation which will help with level 2 plus.

James: 15:47

So being able to take your hands off the wheel, but for watching the road. But that's, again, that's for cars and it it takes especially for level 3, it takes individual countries within Europe to adopt it. And we don't see anything like that happening for trucks at the moment. What we do see is some increased type certification around the technologies that they have to have. So, for example, needing to have automatic emergency braking systems and lane depart systems, that being a requirement of all trucks.

James: 16:19

So we're seeing some progress there in terms of what's the the minimum that you can put on the road. I mean, in in the US, I have to check if they're applied to trucks, but, I think it's 27 or 29. Nitsa has said, all vehicles need to have nighttime automatic emergency braking that protects pedestrians and cyclists. We see those kind of regulations, but I don't see anything at a top level at the minute that's going to help widespread autonomous truck deployment.

Markham: 16:49

In China, the the national government has identified advanced technology manufacturing as a big driver of its economic growth over the in the foreseeable future. And this autonomous trucking would certainly seem to, fall into that category, and, you know, China, you know, is right now is the subject of a lot of, controversy because, of course, they've used very aggressive industrial policy and with, fairly sig significant subsidies to the to the manufacturers. So is this an industry that is ripe for China's type of strategy where, you know, the the government, subsidizes and supports, the manufacturing sector and scales it up really rapidly, and and and then they start exporting and, you know, they're driving down costs, and next thing you know, the, you know, China dominates the the the global industry like they do with solar and batteries and so on. It it might we see a strategy coming out of China like that for autonomous trucking?

James: 18:00

We're definitely seeing that for the autonomous cars market in terms of the the progress that they're making, it being subsidized and heavily aided. I haven't heard the same thing happening for trucks. I think at the minute, China is the the Chinese trucking autonomous trucking companies are working on getting their, technologies deployed internally in a in a way that makes sense. So focusing on safety, particularly in China. So when we did research, we found that per hunt, yeah, per 100,000,000 kilometers of truck driving, China has 37 times more, collisions than the US.

James: 18:51

There's a lot of different reasons about that, why that might be the case. But certainly investing in in safer trucks and autonomy to make trucks safer will be a priority there just to make their roads safer. It also has similar issues to the US in terms of the scale of the country and having long routes to do and being able to cut down on needing 2 drivers in the vehicle. So I think it's got those things that it wants to solve for its own economy as well before it starts shifting the technology elsewhere. And we do see some companies working between China and the US, but, I mean, too simple comes to mind and they ended up in in quite a bit of difficulty and have scaled back a lot.

James: 19:29

So there's always if you do work between China and the US, you you have to be careful.

Markham: 19:35

Final question, James. Are we seeing different business models, emerge in the autonomous trucking sector?

James: 19:44

Yeah. There's a a few different, business models in it. Different companies have have gone down different routes. I would say what we tend to see at the moment because they're in the development stages is the the technology companies and the startups are owning the fleet and being in charge of their own, essentially trying to, like, running their own, logistics company. So they have customers.

James: 20:15

They've got control of the fleet. They're doing doing the roofs. That's what they're doing for that testing. We know that some of them want to go down the route of having a sort of entirely software so that they have a network of roads that you can use their self driving technology on. And then they have agreements with OEMs to build compatible trucks.

James: 20:36

And then if you're operating a truck on that road, you can pay a per mile rate to use the the driverless testing. Oh, sorry, the driverless service. So that's that's kind of one way we that's kind of the driver as a service approach. Some might try and go down the route of of being a logistics company and just setting up their own logistics company, holding their own fleet. You know, it and then the the other option is that they just sell the entire product to an OEM, including the network as sort of per vehicle licensing structure.

James: 21:10

I think we're it'd be exciting to see what happens as it as it evolves. If we look over to the robo taxis, which are slightly more advanced in this route, Waymo and Cruise and everyone is owning their own fleet at the moment. But, you know, Tesla's announcement last week, it seems the intention is to sell robo taxis to members of the public, and then members of the public are their own fleet owners in the same way as as, Uber has scaled. And then we also see that Waymo has partnered with Uber for its deployments in Austin and Atlanta. And Uber will be the only way that you can access Waymo in those places.

James: 21:51

So, yeah, maybe robotaxis are going to go down that route of, the technology providers to do the mapping and get everything up and running, and then consumers are going to be the the fleet owners.

Markham: 22:05

Interesting. Well, James, thank you very much for this is, quite in it's an interesting sector. One that, doesn't get a lot of public attention, and not a lot of information available on it. And so we appreciate your insights. Thank you very much.

James: 22:22

Oh, thank you for having me. Always a pleasure to talk to you.