Art and Science of Autonomous Vehicles (Richard Bishop - Bishop Consulting)

Today Richard Bishop joins the Mile Marker podcast to talk about the art of autonomous vehicles, the benefits of going driverless, and what the future of fleet operations looks like with AVs. 

https://ridecell.com/
https://www.linkedin.com/company/ridecell/
https://twitter.com/ridecell

Transcript

Copy Transcript Copied

Announcer:

Welcome to the Mile Marker Podcast where we explore fleet automation and shared mobility, focusing on innovation for businesses with fleets.

Angela Simoes:

I am your host, Angela Simos, and joining us today is Richard Bishop, principal at Bishop Consulting. Richard's extensive knowledge of connected and automated vehicles lends him the ability to provide strategic counsel with a global perspective to his extensive roster of clients and federal government agencies, vehicle manufacturers, suppliers, research laboratories, state departments of transportation and technology firms. He is a contributor to forbes.com, authoring articles on the impact of autonomous vehicles on fleets, the world, and the bottom line. Today he joins us to talk about the art of autonomous vehicles, the benefits of going driverless, and what the future of fleet operations looks like with AVs. Also, joining me as my co-host is Mark Thomas, AVP of Strategic Alliances at Ridecell. Richard and Mark, thank you for joining me. It's great to have you both here.

Richard Bishop:

Good to be here. Thanks, Angela.

Angela Simoes:

Excellent. So Richard, before we jump into the meaty topic, give us a little bit of background on yourself and how you came to focus on trucking.

Richard Bishop:

Sure. I'll try to make a long story pretty short. I rode in my first automated vehicle back in 1991. Not long after I'd taken a US Department of Transportation position focused on automated driving, and I led that program. It was very car-oriented, but we had some truck aspects as well. I jumped out and started my consultancy in 97 and very involved across automotive ADAS, and other things. But it was all about the urban challenge where I led a team there and the subsequent investment by Google and starting things up that opened up the startup world to hardware, not just software. Vehicles, they're real 3D objects, not just bits and bytes. And as it turned out, the truck side jumped out pretty quickly. I was involved in truck platooning quite a bit in the early days, and then we saw the other single truck full automation systems come into play, companies really cranking up. So back in those frothy years, I was very involved and still tracking the industry very closely.

Angela Simoes:

Wow. 91. I mean, it sounds so long ago. I can imagine just the experience of how far things have come with autonomous vehicles, right?

Richard Bishop:

Oh yeah, yeah. It's kind of freaky. Yeah, sometimes I sound like a dinosaur, but that's-

Angela Simoes:

Not a dinosaur. Just well-versed. Well-versed. So I think the question that nearly every fleet-based business is asking themselves is, why should I consider autonomous vehicles? And then of course then the next question after that is, how do I get started? So what do you say to those companies that you encounter that are asking that question?

Richard Bishop:

Yeah, there's a lot of layers to this. Everywhere I go, whether it's on-road, long haul, short haul, even agricultural hauling, driver shortage is just everywhere. So you have the option to move that freight without a human driving the vehicle is just sort of a survival approach, just you have to move that freight. But there's also so much more than that. It's the operational safety that comes with it. There's even substantial fuel economy benefits because an automated vehicle, an automated truck will drive as well as or better than your very best driver and their very best five minutes of their shift, and they're not impatient and all of that. And operational flexibilities offered as well. You can run that load over the middle of the night when you'd have even more trouble getting a human driver. There's a lot of factors there.

Angela Simoes:

Those are all really great points. And you mentioned something about the efficiencies and operations. At Ridecell, we talk a lot about fleet automation and applying automation to various aspects of operations to help fleet managers really do more with the resources that they have. And as you said, resources are becoming very limited these days. So how do you see autonomous vehicles fitting into the ability to have fleets be more automated, without completely removing the human subject?

Richard Bishop:

Yeah, there's various elements there. The long haul trucking, the dominant early deployments will be what's called a ramp to ramp activity. So the vehicle is tuned up to run on a limited access freeway, and that's orderly in general, but not take the load all the way to the loading dock. So there's the idea of a transfer yard very close to the interstate entry and exit, where a human driven vehicle brings the load from the distribution center or the manufacturing plant to that transfer yard, and then that trailer is unhooked and rehooked by an automated vehicle. And so it's a very technology-human interaction in that sense.

Angela Simoes:

So you're talking last mile is kind of how the industry refers to that, right?

Richard Bishop:

That's right, yeah.

Angela Simoes:

But I think in some of your research that you've done, and I think you just actually published a paper about the last mile solution and how automation or autonomous vehicles are being applied there as well. Wondering if you see an opportunity where you can connect the two through fleet automation processes or technology.

Richard Bishop:

Indeed. And broadly speaking, there will be ripples, or more than ripples, through the operational world of a truck fleet because of the automation. The timing of when a load goes, whether it goes in a box truck versus a full-size tractor trailer if it's more short haul B2B. There's a lot of truck operations. The larger envelope of automation around truck operations with the AVs will evolve a lot. And when I give talks, I put that out there as a, Hey, keep an eye on this. I don't know how it's going to evolve, but in essence, every element of a truck operation, a truck fleet gets touched by this new tool of automation.

Mark Thomas:

Keying off your last mile. Do you see a use case evolving where the vehicle itself ends up being self-driving, but there is still an attendant in the vehicle to deliver the packages, take the product and get it the last a hundred feet to the person's doorstep? How is that use case evolving? I haven't heard too much about that.

Richard Bishop:

Yeah, that could be the case in a B2C kind of world, residential delivery. And that is not seen as an early use case because it's hard. It's as hard as the robotaxis that are trying to get their act together in San Francisco, and we'll get there. And of course some of the robotaxis companies have partnered with Uber Eats and others to start doing that residential delivery.

But when it comes to freight, and the companies that I'm involved with, that big world is B2B. There's a significant amount of B2B activity and short haul. One of the companies really the only company doing that now is Gatik. And they've been able to simplify the operational design domain so that for Walmart, their first customer, they're running from a Walmart dark store to a Walmart retail store, back and forth. It's the same route all the time. As you can imagine, technically, you can really certify and validate that operation because you have high awareness of that roadway and its moment to moment dynamics, so different than trying to just cover an entire geographic area. And that market is growing very quickly. It's not just about the long haul trucking. The Gatik operation for Walmart is a box truck operation. It can easily be an EV operation, if the fleet wants to do that.

Mark Thomas:

You've been in since I guess 91. 5 years ago, I think we hit an inflection point in terms of the trough of disillusionment. Now we're pulling back out of it. What are the hallmarks of what we were thinking five years ago and what we're thinking now in terms of the use cases that have really defined themselves and to allow this to be back and be relevant.

Richard Bishop:

You're right about that trough of disillusionment. To me, it seems like there's been several of them over the last few years. And yet the industry, of course, any frothy new industry is going to have lots of players early on and then it starts to consolidate. Some folks drop out. That's what's happened recently. And yet there are still some very strong players who are moving to market. Torque and Aurora are very strong, Plus, whereas others are, there's kind of a question mark and what it takes to get from where we are now, which is essentially that the technology is either ready or almost ready.

Well, then what? Well, you have to scale up and scaling up means having the right kind of partners across the industry, whether it's Tier 1s and of course OEMs, that type of thing. You have to be able to handle a more varied environment. Initially we'll see the automated trucks running in more benign environments, but enable, in order to really serve the freight industry, you're going to have to operate across the whole geography of the US and whatever other markets you want to enter. So there's a professionalism that has come in that's beyond these super smart software guys figuring out how to make a truck drive itself to the business level and a whole new world for these companies. And they're stepping up to it.

Angela Simoes:

I want to go back to your Walmart example. Because it's kind of the same route over and over again, they're sort of perfecting the efficiency of that route, I would imagine. And you had touched on some of the cost savings and other efficiencies that trucking companies experience when they go autonomous. Can you talk a little bit more about, let's say the revenue aspect of that and where they're experiencing gains in revenue because of cost savings there?

And then secondarily, the maintenance aspect. Because it's such an efficient loop, if you will, any feedback from them or have you heard their maintenance is more regulated, so it's not as expensive or it's the vehicles know sooner that they need maintenance than a human would? And so the system is able to flag it, and whereas a driver may push it another a hundred miles, the truck knows that it can't go any further. So if you could address those two kinds of things that came up when you talked about that example.

Richard Bishop:

I'll speak to maintenance first and then loop around to revenue. A lot in maintenance. There's certainly the question of who's going to do the care and handling of these vehicles, the radars and the lidars, and make sure everything's working. Early on, I think it'll be the developer companies that have to have a present at their customer's site working that, but that may not be efficient in the long term. That's just to launch an initial operation. So the third-party maintenance players are starting to partner with fleets that are running a Kodiak system, for instance, rider and other companies like that. I think we'll see a lot more of that. There's also, maintenance is closely linked to inspection in the truck world, and in the past there's been a pre-trip inspection and then a cop can pull you over at any time to do a roadside inspection.

That's pretty challenging for an automated vehicle to handle. There's been, boy, I think a three-year process of the developers working with the Commercial Vehicle Safety Alliance, which is the organization that brings together all those state police inspectors to come up with a depot based inspection. It's what buses do. You know, don't want your passengers in your bus to sit on the side of the road while a cop inspects the bus. So there's a depot based inspection before the trip, and there's a lot more detail to that. But the point being that there's a very structured approach to doing that, and you'll need specially trained inspectors within your own operation certified to do the inspections. And that's linked very, very tightly to the maintenance. And as you mentioned, Angela, the potential of knowing ahead of time that something needs attention. These trucks will have so much smarts watching the smarts of the vehicle that you'll see not only a decrease in crashes because of the technology, but a decrease in maintenance, emergency type maintenance activities.

Mark Thomas:

In your world, do you see the people that are maintaining these vehicles going to the vehicle, or does the vehicle need to be brought to the maintenance?

Richard Bishop:

That remains to be seen. It's going to be very company specific, but part of that, what does it take to get to market, is what we have now is the freight carriers. The traditional players are having to adapt their operations for the special aspects of automated driving, such as the transfer yard idea. Well, the competitive advantage will be those who can move that situation back to a place where the fleets don't have to bend over backward to use this technology. So the maintenance side is a piece of that, what works best for the fleet in an efficiency sort of way, in a profit sort of way. And it remains to be seen.

Mark Thomas:

Are you seeing the autonomous vehicles autonomously driving themselves to the maintenance places?

Richard Bishop:

Ah, nobody's ever asked me that question. That's-

Angela Simoes:

They're smart. Are they that smart?

Richard Bishop:

My guess is early on, a driver will have to be dispatched out there. There'll be some sort of a whole new business happening where companies can be available to intercept a self-driving truck in the middle of Arizona or something that's on the side of the road and drive it to where it needs to be. If that happens very often, the industry's not going to be in a good place. But you have to have-

Mark Thomas:

Even routine maintenance, though, the vehicles themselves, when they're not broken down, can move themselves about to take care. And when you're in a situation when there is no driver, how do you then handle keys? Does everybody essentially use their phone to access the inside of the vehicle, the cabin, and get permission to drive it?

Richard Bishop:

That's a great question, and I think each company's handling it different, but I don't know how that's being done. I would expect it's a highly digital sort of thing, phone and whatever.

Angela Simoes:

Great. And then there was the revenue side of the question. Will you address that?

Richard Bishop:

Yeah, sure. That's the key thing. These days, these companies, they're either publicly traded or they're venture capital supported, most of them, and they need to get that revenue really cranked up, which means they need to deploy with their automated truck. I'll go back to Gatik because they're the only company that has deployed the commercial product and they have their trucks are moving driverlessly out in the real world. They're doing that in Bentonville, Arkansas. They're doing that in Toronto with Loblaws, a major retailer there. And soon they'll be doing it in for several customers in the Dallas Fort Worth area. The customers have been announced and they're on a process just to go through the paces to approve the driverless operations. And what we're talking about with Gatik and other companies who will enter this space is even working with the customer. So the customer has a sense that their cost of operations is going to go down even while Gatik's revenue is going up to a good place.

And those kinds of business arrangements are at a mature point, and we'll see more of that. But it's all about scaling up the numbers and the numbers of vehicles, numbers of miles. And you can only do that if you have a truck manufacturer who's ready to provide to you what we call a Level 4 ready truck, a truck that's got additional redundancy and brakes, steering, even the electrical system and the cooling system to support a driverless truck. So that'll sort of pace the revenue generation possibilities for these companies. Right now they're doing it with dozens of trucks or maybe less than that. And that's good enough to demonstrate to customers that the system will fit in operations, but there'll be an inflection point when the truck OEMs, the Class 8 truck OEMs can deliver those. Meanwhile, if you're running smaller vehicles such as the box trucks, that's much more easily done, the redundancy. And that's already been done.

Angela Simoes:

I feel like we can't have any technology conversation these days without mentioning AI. And some of the things that you're talking about it... Again, this sort of feedback loop and learning from the roots and learning from how the vehicles are performing. And then think about, you mentioned, all the sensors, the lidar and the cameras, and these are all feeding back into the system.

So have you heard of companies that are taking all of that data and applying some AI... Or maybe not AI, maybe it's another way to do it, but learning from all of that data to then make the roots or how the vehicle operates more efficient, and whether that be fuel economy or maybe it's a different route because it's shorter or anything like that, but again, would be remiss if we didn't mention AI and how the vehicles and the systems are getting smarter with every route that they drive.

Richard Bishop:

Great question. I'll have to restrain myself from going off on a rant about-

Angela Simoes:

We can have you back for another podcast. We could talk about it.

Richard Bishop:

Well, the term has been corrupted in a way because the society now hears AI, they're hearing about generative AI, ChatGPT, and all that societal implications. But in their minds, they're thinking AI and only AI, whereas we know AI has been around for years and it's the reason we can talk into our phones and it'll write a text for us and all this stuff. There's sort of basic AI and then there's this form, fantasy stuff. So the development of automated driving relies deeply on the AI processes of ingesting a lot of data, which can help.

One example would be computer vision and being able to detect and understand what's happening as a car moves around on the road, pedestrians and shadowing and all that. But see those effectively. And one company I know of five or six years ago, they went to dash cam videos that are on YouTube, a huge number of those, and they just shovel that in to their AI engine. And it got very smart, very quick because it had that level of data. And the truck guys are doing that as well. One reason they're doing so many drives with safety drivers is they're exposing the vehicle to the real world, ingesting all that information and tuning up their ability to perceive the environment, it's just a given in this space.

Angela Simoes:

Excellent. All right, so we are coming up on the end of our podcast. It's just another question or two. So everybody's trying to figure out, and people have made predictions in the past, we will have full autonomy by X date and everybody was wrong. But when you look at fleets or fleet based businesses fully adopting autonomous vehicles, when in your opinion, when do you think that might happen and what will it take for fleets to fully adopt? I mean, it's not like you just swap out gas powered vehicles. So what will it take and when do you think that might happen?

Richard Bishop:

I'll give you a couple of examples. Last week I was at a conference in San Francisco and rode in both the Waymo and Cruise robotaxi vehicles in very gnarly traffic. And it was just flawless. It did great. This is a commercially available technology. It depends on where you are geographically, but it's commercially available. And that's on the passenger carrying side. On the truck side, the freight side, as I said before, one company, Gatik has already entered into that driverless world. I think they were the first country company in the world to start doing driverless freight.

The other guys have said 2024, it'll be announced, it'll be released commercially. And these companies, that's Aurora and Kodiak. Aurora in particular is working with some truck manufacturers to provide that Level 4 redundancy I spoke of. So the pieces are coming into place. And another strong player being Torque Robotics, which is 51% owned by Daimler trucks, they're cranking it all along very solidly. So it's not far off. And depending on where you look, it's already here. It's happening.

Angela Simoes:

And the thing about that is that the passenger side of things will be able to learn a great deal from the trucking side, just in terms of some of those use cases of how do you navigate traffic. I realize highway driving is much different than urban driving, but there's still, I think a lot of learning to be done and system improvement.

Richard Bishop:

And you haven't mentioned Waymo yet, but they have both a very strong robotaxi program as well as a truck automation program. The same is true of Aurora, robotaxi and truck. Currently, Waymo is focusing virtually all of their energy on the robotaxi side. They've got to ramp that up with the revenue. They've got to get profitable. So they put their truck side on ice for a while, but they'll be very strong once they step back in. And Aurora's the opposite way, so they're putting all their energy into the truck side, but that allows them to pivot quickly on street trucking later and their robotaxi operations.

Angela Simoes:

Excellent. Well, this has been quite fascinating. I learned quite a bit about what's happening in the trucking industry. I thought I was up to speed on everything, but I learned a few things. Any final thoughts you'd like to share with our listeners?

Richard Bishop:

Simply that while you can find lots of naysayers in media articles, it's a very strong and very mature industry, a very responsible industry. And we will see the safety promises. I'm certain we will see the safety promises become reality and the whole society will benefit from that, even while freight flows more cheaply and more efficiently.

Angela Simoes:

And Mark, any thoughts on your end?

Mark Thomas:

I agree. This has been really informative. Thank you so much, Richard.

Richard Bishop:

You're welcome.

Angela Simoes:

Yeah, this has been great. Well, thank you both again for your time. Thank you to our listeners who joined us for this episode. And until next time, keep moving forward.

Announcer:

Thanks for listening to the Mile Marker podcast. If you liked what you heard today, give us a like, share this episode on social media, and even take a minute to leave us a review on iTunes or wherever you listen to your podcasts. Be sure to join us for our future episodes full of insights and ideas to keep the mobility industry moving forward.