AEye Insights: The Roadmap to Autonomy
In the first installment of the AEye Insights series, AEye Advisor, Jim Robnett, interviews Elliot Garbus, AEye Advisory Board Member and former VP & GM at Intel Transportation Solutions Division. Elliot outlines some of the technical challenges facing the industry today, what kind of partnerships we’ll see in the future, the social and legal issues related to autonomy, and where he thinks transportation is headed this decade.
AEye: Welcome to AEye Insights, where we talk trends with proven industry leaders. My name is Jim Robnett. AEye is a leading provider in the robotic perception world. We’re working in the automotive markets as well as other transportation markets, including trucking and rail. With me today is Elliot Garbus. Elliot is a board advisory board member for AEye and former VP and general manager of the Transportation Solutions Division at Intel. Elliot, welcome.
EG: Thank you, Jim. Happy to be here.
AEye: You have a really interesting history with Intel. Can you share a snapshot of what you’ve worked on in your past?
EG: Just to be brief so we can really get on to the meat of the discussion, I spent almost 30 years at Intel, my last role as the VP and general manager of the Transportation Solutions Division. I got a really up-close view of some of the transitions that the automotive industry is wrestling with. You’ve got an industry where the major players are about 100 years old, and that’s a stunning accomplishment for any company to have survived for one hundred years, and now they’re facing a number of rather dramatic transitions: the shift to electric propulsion systems and the rise of both active safety and ultimately autonomous vehicles, all of which are tremendously disruptive and require a tremendous amount of agility.
AEye: Let me start with an easy question for you. I think you and I are about the same age. We grew up in probably similar circumstances. We were promised jetpacks and self-driving cars. So, Elliott, which are you going to own first, the jetpack or the self-driving car?
EG: Well, I think I probably won’t have a jetpack and I probably won’t own a self-driving car, but I’ll be happy to ride in them.
AEye: So, Elliot, let’s start with just a snapshot. Give us your view of autonomy, because that’s what we’re really here to talk about. Obviously, electrification and connectivity and shared ownership, to your point, is all part of that. But tell us how you see the autonomous vehicle industry from a broad view, and then we can get into some of the technical issues that we’re facing as an industry and maybe some of the social and legal and other issues that we’re facing as an industry.
EG: As I talk about autonomy, I’m going to focus first on Level 4: the driverless vehicle, just so there’s no ambiguity. And in many ways, I think this has the opportunity to be the most disruptive. First, if we think about costs today, if you were to take a taxi, it costs about $6 a passenger mile. If you take a ride share, it’s about $2.25 a passenger mile. If you drive yourself, it’s about $1/$1.75. Bloomberg estimated that by 2025 the cost of riding in a fully autonomous vehicle – a driverless taxi – would be $.35 a passenger. Today, mass transit is $.30 per passenger mile. So, you’re talking about the convenience of point-to-point transportation at a cost that approaches mass transit. And I’ve spoken with some auto executives who have been firmly convinced that they will get there, with their operating costs below $.35 per mile, so these are levels that really set up fundamental and dramatic industry-level disruption.
AEye: But if we go back two or three years or four years ago, what we heard from a lot of those industry executives was we were going to have those vehicles on the road today. What’s holding us back right now?
EG: I think a lot of it is concerns about technology readiness, and I think most of those companies were making announcements around 2021. They were being aggressive. They were setting tough to reach goals to inspire their teams to deliver. I think in many ways there are a tremendous amount of unknowns here. We do see some rather remarkable milestones in this space: Waymo has announced here in the suburbs of Phoenix that they’re now operating their autonomous vehicles without safety drivers and they’ve opened up their platform for any user to join within their geo-fenced region. Aptiv has a joint venture (JV) with Hyundai Motional that is partnered with Lyft and they have permits in Las Vegas to operate the vehicles without safety drivers. So, we’re seeing the very early phases of this transition, and I think there are a couple of things that are important to recognize. One, you do have a tremendous amount of technology, and responsible companies want to make sure that they are delivering better than human levels of safety as they roll these technologies out, so I certainly applaud that thinking and direction. But the other issue is there are a whole set of customer experience issues that need to be addressed when you’ve got a driverless vehicle that are not always obvious. Just a couple of silly examples: If you’re in a fully autonomous vehicle and a passenger that leaves the door open, is the vehicle disabled? It’s not like a driver can pop out and close the door, so it changes the requirements about the kind of vehicle. Waymo in their pilots has discovered it’s very important that they have the ability to wake up passengers because people fall asleep in the vehicles all of the time. I think how people want to use the technology, what they need to get comfortable with it, and how you build trust between a passenger and a robotic vehicle – there’s still things to learn and work out before you can roll a business out at scale.
AEye: So, let’s start with some of the technical issues. If we think about a fully autonomous Level 4 driving stack, we start with sensing, we get to perception, we get to driveable area and then path planning and then the robotic driver. Maybe that’s a way to look at this stack. From a purely technical perspective, to deliver that safe driving experience, where should we be spending our time and where is kind of the value being created in terms of delivering that experience?
EG: I think there’s still a lot of innovation going on at almost every level of the stack. We see millimeter wave radar, increasing levels of accuracy – certainly what AEye has been doing in the combination of LiDAR and camera has opened up innovative uses of LiDAR with much greater levels of efficiency. These are all remarkable, and cameras continue to get better dynamic range, better color fidelity – all nice improvements in the sensing modalities. As we talk about perception, there’s certainly been a lot of enthusiasm around deep learning. Deep learning leverages neural networks. It is a relatively new technology having really emerged around 2012-13 and initially used for image recognition, which is computer vision, which is the role it’s playing in vehicles. And we see a lot of work going on in the semiconductor industry to improve the performance and performance per watt of the technologies that deliver on deep learning. But I think for getting us there on autonomous driving, these technology improvements I’ve just described are all nice, but they’re not necessarily necessary to get us there. Those modalities are sufficient today. The improvements are certainly welcomed. I think the challenge is capturing enough of the odd corner cases that humans run into driving that you need to be able to train a computer on. And because those corner cases tend to be rare, it requires a lot of time to capture them.
AEye: Right as we solve these technical challenges, the business challenges and maybe the social legal challenges need to be solved along with that. And you talked about how companies view their cost targets and there maybe even their revenue targets in the future. From a business standpoint, what do you see in terms of emerging partnerships or emerging business models? What’s changing on the business side, along with the technology, as we move these things forward?
EG: I think there are a number of things and a number of industries are going to be significantly impacted. It’s estimated that once we get to really wide deployment of fully autonomous vehicles, the average American family will save $5,000/year. You won’t have auto insurance to pay for, you won’t be buying cars or making car payments. That’s really a stunning amount of savings. There’s going to be a tremendous reduction in injuries and the loss of life. Today globally, we see about 1.3 million people a year dying in traffic collisions. In the US, it’s about 40,000 people a year. That’s a shocking number, right? 40,000 a year is like a 737 falling out of the sky every weekday. So, having those numbers go from 40,000 people to zero, that’s remarkable, that’s something to aspire to. In the US, there’s an additional 50 million people who are injured. And again, we have a real opportunity to create some strong benefits for society. But from a business perspective, I think one of the things that’s most fascinating is who is actually going to pay for the ride? You’ve got an opportunity now, particularly as the cost of transportation comes down, to find someone else to pay for the ride: you make a dinner reservation with your wife and the restaurant decides they’re going to throw in transportation, and maybe you’ll decide to buy an extra cocktail or bottle of wine and they would have ended up ahead. Or you’re going to Target or Whole Foods and the retailer will pay for the trip. So, a lot of very interesting things are going to happen, and these kinds of partnerships and the ability to build an ecosystem of people who are willing to pay for a trip and match them up with passengers is going to be a critical piece of the business.
AEye: That’s really interesting and I think, like you said, really disruptive. And there’s opportunity and challenge on that side. For the people who don’t move fast enough and don’t adapt, there is the real risk of being left behind here. And I think that fear and opportunity are some of the dynamics that we’re seeing across all of our customers right now.
EG: Certainly. It was very interesting as I was out visiting with OEMs around 2016, we actually saw all of the OEMs make a rather dramatic shift in their thinking, and some of the fear you mentioned was looming large. I had one of the world’s largest OEMs express concern that they could become Kodak. A huge, enormous, very well known, almost a beloved brand, that the transition to digital photography pretty much took out, and that these trends in autonomous driving could really destroy a lot of the auto industry as we know it. It was also an observation by another very large automaker that Google could buy any car company they wanted and there was not a car company that could consider buying Google, which had one automaker say to me, you know, we don’t want to become a Google subsidiary.
AEye: For some of us old guys to be in that new reality, it’s pretty amazing.
EG: I think, again, inside a company that’s 100 years old where the structures have really been hyper optimized for cost – let’s face it, automobiles are extraordinarily complex consumer products that are heavily regulated and delivered with extraordinary levels of quality. I mean, it is really, really a remarkable achievement, and these 100 year old companies are highly optimized to deliver these high quality products in a cost effective way, and suddenly having to wrestle with all of these transformations is really challenging, and you can see how different companies have addressed it. It’s fascinating to see General Motors, for example, they acquire a small company called Cruise, they put long time GM executives in charge, they step it up – they place an enormous bet on autonomous driving. Delphi, a Tier one supplier of historic standing, splits the company in two, and creates Aptiv to focus on active safety and autonomous driving and they end up in a JV with Hyundai. Remarkable changes in the industry, and I think this is really all about, as you rightly said, balancing the opportunity and the fear.
AEye: We’ve talked a little bit about the history and where we are now. Do me a favor and pull out your crystal ball and tell us where we’re going to be 5 to 7 years from now. We’ve got this kind of discrete world of mobility and this discrete world of ADAS or traditional automotive. I think we might agree that those are converging in a way. But what’s the day in the life going to look like for transportation 5 to 7 years from now?
EG: Well, let’s talk a little bit about the near term, because I think in many ways, we’ve passed that by. I think there are a set of remarkable things going on today as the automakers largely have embraced active safety, and in particular, I think Toyota’s demonstrated remarkable leadership with their Toyota safety sets where they’re integrating forward-facing cameras and a very cost-effective suite of simple ADAS features from the lowest end vehicles up to their highest vehicles. I think that’s just remarkable. And the opportunity, again, for these technologies to really reduce injuries and save lives, we’re going to see a greater and greater embrace of those. We’re beginning to see the adoption of Level 3 technology – I like to refer to these as scenario-based driving assistance. We see General Motors has done some remarkable things with their Super Cruise, and certainly Tesla has done some remarkable things in this space as well. We’re going to see more of this, where automakers roll out features, where the vehicle can drive itself on the highway or the vehicle has a Traffic Jam Assist, and if you’re in bumper to bumper traffic, you can just set the car to operate itself.
There are some automakers that at least at one time have called out some of the problems with Level 3 automation and safety. Level 3 SAE Automation requires that the driver stay in the loop, be ready to take over control of the vehicle. Ford announced some years ago that they were very concerned because they couldn’t seem to keep their engineers engaged to take over the vehicle when the vehicle needed a driver to intercede, and so I think we’ll see more and more of that sort of scenario-based driving come in, more and more active safety come in. But the approach you need to take and the business model between active safety and full autonomy is huge. And, when you move to a driverless vehicle, it really makes economic sense, as I called out earlier, for that to be a – transportation as a service – and that brings on a complete and disruptive change in business model.
I think in the near term, we’ll continue to see continued enhancements and embrace around active safety, making our vehicles safer, and then, as we look out a little further in time, we’ll really move to fully autonomous vehicles that will embody a much different business model. I spoke with one of the leaders in ride sharing some time ago, and they said, if they could get a vehicle that costs $300,000 and had a lifetime of only 2 to 3 years, they could make the business model work because they’d have a vehicle that’s operating 80, 90 percent of the time, whereas today we buy a vehicle and we operate it at best four percent of the time. Some people joke we buy cars for the right to park them. It’s one thing to think in a very U.S.-centric way. If we think globally, the U.N. has estimated that by 2030, two thirds of the world will be living in cities of ten million or more people, and in those megacities, you have horrible problems with air pollution, with traffic and with parking, and while we’ve talked about some of the disruption in transportation. If you think about cities for a minute, about 30 percent of the real estate in cities is used for parking, and so there’s an opportunity to free that real estate up for much more effective uses. I think globally, we’re at the beginning of what’s going to be a really remarkable transformation that certainly in 5 to 7 years I’d expect we’d see broadly deployed.
AEye: It’s an incredible time. It’s an exciting time in a lot of ways. This is our industry’s moonshot and it’s fun and pretty amazing to be a part of. Elliot, thank you for joining us. Really appreciate your insights. You’re a great asset to AEye as one of our key advisors. We appreciate your time as always. Thank you, everybody, for joining us for the first edition of Insights, and we will see you next time.
EG: Thank you, Jim. And stay healthy.