Autonomous vehicle (AV) design is complex, to say the least…
The complexities of autonomous vehicles
In the past, designing a vehicle required addressing issues such as speed, handling, safety, and efficiency, among others.
While this is complicated, it is much easier compared to what it’ll take to build autonomous vehicles.
Not only do you have to take care of all the vehicle design challenges, but you also have to design its ecosystem. You must consider how it will connect with the infrastructure, its users, and the controller. Effectively combining all of these needs requires the collaboration of engineers from different disciplines.
In this episode, the second part of two, Ed Bernardon interviews Felix Andlauer, VP of Mobility Solutions at NEVS – a Swedish electric and autonomous vehicle (AV) manufacturer founded in 2012 with the vision of shaping mobility for a sustainable future. He’ll help us understand what it’ll take to adopt AVs fully.
Some Questions I Ask:
- How far away do you think we are from having an autonomous vehicle? (01:43)
- What have you done to the vehicle to protect the passenger? (07:25)
- Do you think there will be a standard language for AVs? (20:13)
- Why do you have a remote driver system? (21:39)
What You’ll Learn in this Episode:
- How a NEVS car safety design differs from other AVs (04:55)
- How NEVS handles the range problem (08:22)
- What it’ll take to create a smart mobility ecosystem (15:42)
- How to make engineers work together (25:17)
Connect with Felix Andlauer:
Connect with Ed Bernardon:
Ed Bernardon: How has the development of autonomous vehicles changed over the past few years? How will AVs improve traffic congestion, available urban space, and sustainability? How does constraining what is expected of AVs make their realization occur sooner rather than later so we can get benefits from AVs for more affordable transportation, reduced traffic, and need for parking space? What other measures are needed so you will feel comfortable in an AV? For instance, remote intervention for safety or an adaptive interior to maximize privacy? According to my guest, this will happen in the very near future and it will be fully electrified to help us also meet our sustainability goals. NEVS is an electric car manufacturer, based in Sweden, whose company goal is to build mobility ecosystems aimed at a sustainable future. The company aligns itself with trends such as connected vehicles, electrification, changing ownership models, as well as autonomous driving. NEVS’ PONS mobility system includes autonomous vehicles, a connected user interface, and full fleet management and service. Sango, the AV at the core of this service, has been optimized for shared mobility with adaptable privacy modes. This paints a new and unique picture of what a future AV experience will be.
Ed Bernardon: Welcome back to Part 2 of my interview with NEVS’ VP of Mobility Solutions, Felix Andlauer. Last week, Felix and I discussed the current state of the AV industry, and what NEVs is doing to stand out from its competitors. This week, we continue our discussion as we cover topics such as optimizing EV shuttle range, flexible fleet services, regulation and standardization, remote driving, and third-party collaboration. Join me, Ed Bernardon, on this episode of The Future Car as we imagine a future with AVs.
Ed Bernardon: How far away do you think we are from having an autonomous vehicle? Well, firstly, it would have to become commonplace just doing the basics that we’ve been talking about driving people around. When do you think systems like yours will be commonplace? And how much farther down the road after that would be this opening up of the platform so that other startups could take advantage of? How do you see that evolving over time?
Felix Andlauer: I think the first step to getting vehicles out on the road so that they, without a safety driver, in the vehicle, can provide services, in very confined areas where we can control the safety case and we can know we can handle all situations, we think this will happen in about two years.
Ed Bernardon: And what do you mean by “confined area” in this particular case?
Felix Andlauer: It’s public roads, no fence or anything around it. But in Europe, in order to get a permit to drive around, you have to define an ODD (Operational Design Domain) in which you can claim that everything that could reasonably occur in this area, we can handle and so we’re safe. The vehicle wouldn’t be allowed to drive out of this area autonomously. And that, of course, also means we can only provide mobility needs that start and end in that area. But this we think will be possible really soon. And I think that’s also to your initial question about what’s happening in the AV industry. I think a lot of companies are realizing now that this AV Level 5, so you can drive everywhere safely, is really far away. And I really agree but as soon as you limit the area that you operate in, then it becomes much easier. So, we think that providing a service in a defined area will be possible in two years. From then on, fleets will grow, areas of operation will grow, and use cases of how people use that kind of service will grow over time. But when we reach that point where in the whole city, I can drive autonomously and I don’t need any safety driver, then it will ramp up super fast because it will be cheaper and better than the existing solution. And this is the moment where you get this disruptive change. And from then on, as soon as there are autonomous vehicles driving around everywhere, I would say that this disruptive change will happen in maybe 2027 or 2028. And from then on, I think that the step to it becoming an open platform with a lot of different companies integrating their services goes really fast.
Ed Bernardon: In a couple of years, 2025-2026, something like that, now we’re going to start to see them. You’re going to probably go in the city and expect to see an autonomous car here or there or a shuttle. Then another two or three years after that, where you finally hit that critical mass, and now it’s going to start to become commonplace. So, somewhere around 2030, maybe that’s when we start to see the true power of autonomous vehicles coming in where you can start to do some of the things that that we’ve been talking about. Your vehicle, low and high speed, obviously, your fleet is high speed, it can also be low speed. So, city and highway, and I think when you take a vehicle and you want to make it capable of working at highway speeds, there’s a lot more you have to do to it than if it’s 25 miles per hour, 40 kilometers per hour or less. So, what are some of the challenges that you have to address, especially from a crashworthy standpoint, say, for an autonomous shuttle of your type that’s capable of going on the highway?
Felix Andlauer: So, for us, we’re started out of the bankruptcy of Saab Automotive. So a lot of our people are former Saab engineers, and all our legacy is around passenger cars. So, for us, making a vehicle crashworthy for high speeds is no problem at all. We see that, of course, a lot of other companies in that field are startup companies without this automotive legacy. And then, of course, it’s really hard to get to that point. But for us, designing the vehicle for high speeds isn’t that much of a challenge, really, not more than for a passenger car. Then, of course, for the AV system, it’s still challenging. So we design our vehicle so that the hardware is capable of driving 110 kilometers per hour, but it will probably take some time until the autonomous driving system is good enough to really do that. But we want to have the hardware capable of it, and then we can update the software when it’s ready.
Ed Bernardon: So, city streets first at a lower speed, then eventually getting out on the highways.
Felix Andlauer: And again, the logic is the same, we want to replace the privately-owned car and no one owns a car that can only drive 25 kilometers per hour. So if you really want to replace it, you need to be able to do the same things. So at some point, we need to be able to drive on a highway. In many cases, from an autonomous driving perspective, the highway is actually not very hard. There are not many difficult decisions to be made on a highway.
Ed Bernardon: When you have an electric vehicle, we’ve seen some of these examples of crashes on highways where suddenly the batteries catch on fire. So if you’re going at high speed, the design has to take that into account, I would imagine, which is something you don’t have to do as much or consider as much when you’re going at a lower speed. What have you done to the vehicle to not only protect the passenger, but to protect all the batteries?
Felix Andlauer: I’m really not an expert in that area, but I wouldn’t say that it’s any different from any other electric vehicle that goes at high speed. Also, I don’t have any statistics to back that up, but my impression would be that electric vehicles are probably catching fire more seldomly or less often than conventional ICE cars, it’s just something you don’t read about in the newspaper. So, I don’t see vehicles catching fire to be a fundamental problem, no.
Ed Bernardon: It’s a problem that all electric vehicle companies are having, be they autonomous or not autonomous. But if your electric vehicle is on the road all the time, the range has to be a concern. Everyone worries about range. How do you handle the range problem?
Felix Andlauer: Well, it’s actually a very interesting thing that we run a lot of simulations and we do a lot of calculations. And what turns out to be something that I didn’t expect was that the range you need in a fleet of autonomous vehicles is actually much smaller than in a privately-owned car. And the reason is that in your own car, everything is pretty much the worst-case requirement — you have 500 kilometers of range because you drive 500 kilometers once a month. You need super fast charging because you use that every second week or something like that. Whereas if you have a fleet of vehicles, it’s much cheaper to have 5% more vehicles in your fleet, and then always let one of the vehicles charge and the rest are taking over, than building super big batteries in every single vehicle. Also, from a user perspective, in your own car, the worst case is you run out of electricity or the battery is empty and you stop. Whereas in our case, the worst case is the car you sit in needs to go charging so it stops, another one pulls to the side, you simply switch vehicles, and then you go on driving. So you lose maybe a minute or two, and that’s your worst-case scenario. So it’s actually something we didn’t think of before we started, but that became obvious when we did the calculations.
Ed Bernardon: I think that’s a great example of designing a system rather than designing a vehicle. Because if you’re designing the vehicle, like you said, I have to design range into that vehicle. And trying to do that can make it heavier, can make it more expensive. Whereas, you really just need to provide a range for the customer.
Felix Andlauer: And it’s not only the battery, there are so many things like that. How many people play golf once a year; therefore, they need a large car where they can put the golf bag in. It would be much cheaper for them to rent a car this one time a year with a large trunk. But yeah, you always buy your car for all your combined worst cases.
Ed Bernardon: Can you configure your vehicle, say, I went two people with a large luggage compartment? Is that a possibility?
Felix Andlauer: Yes. But we thought that it’s actually much easier — simply, there are just two options: you book a seat or you book the vehicle. Already, when you’re two people, simply book the whole vehicle because sharing with strangers, if everyone is on their own, is much nicer than sharing with strangers when two people in the car know each other. So the two options are really “book a seat” or “book a car.” And if you book the whole vehicle, then, of course, you can bring your golf bag as well because it’s yours for the time you’re using it.
Ed Bernardon: So, what exactly is the product and who do you sell it to? Like you said, you’re not selling individual cars to people. So, what is it you’re selling? Is that a service? Is that a product? Is it both? Who do you sell it to? And what support do you provide afterward?
Felix Andlauer: So the answer to your either-or question is yes, we’re doing all of that. So, the idea is that if you develop the whole service, then we already know today that the markets will be different. So, just in our direct neighborhood, in Norway, the city of Oslo is very advanced in Europe, when it comes to autonomous vehicles. And there, it’s already pretty clear that they see this kind of service as part of public transport. So they define the role of public transport not only to be operating buses and trains but also smaller vehicles. So, in this case, we would provide the whole service, but the user interface will be through the public transport authority in Oslo called Ruter, so it will be through their app and we can only provide the fleet of vehicles and the operations and so on.
Ed Bernardon: You would run the vehicle in that case. In that example, NEVS is operating the system as well as developing it and manufacturing it.
Felix Andlauer: Not necessarily. We could, but we could also partner up with the company that already operates buses and trains because they already know how to manage workshops, so we don’t need to do everything ourselves. This holistic approach that we try to develop everything is not that we have to sell everything but that we want to look at the system as a whole and we always have a fallback. So if we don’t have a good partner to do it, we want to be capable of doing it ourselves. But in many cases, there’s already someone that is better at it. If there’s a country that simply says, “Yeah, sell the service directly to the end customer like Uber does it,” then we can do everything ourselves. But whenever someone else wants to do a part, for the overall business, it’s better to let them do it, then we can do that as well. So, one scenario we discuss a lot and I think it’s super interesting is actually corporate customers. So we’re not selling to the end user but we’re not selling to the city either, but we may be selling to a company that we transport the company’s employees or the company’s customers because especially in the beginning when you are in a confined area, you need to find customers that have a mobility need within that area. And then often finding a company that maybe has a bad connection to the next train station could be super interesting because they want their employees to not drive to the office because there are no parking spots, but the employees are not taking the train because the last mile connection is bad. So if we could sell to the company that will solve the last mile connection for their employees, everybody wins. So we already know that there will be different sales channels for that, so we need to build our whole service in a modular way so that you can pick and choose which components to use and which ones to not use.
Ed Bernardon: So, in one extreme, let’s just say, I could say to NEVS, “Hey, come into my city. I want service between the train station and the subway.” And you could run the whole service and say, “Okay, thank you. We love this privilege.” But you make all the money on that, you’re just providing the service. Is that one example?
Felix Andlauer: That could be one example. And I think even already today in public transport, there are different kinds of contracts. Sometimes the public transport operator gets paid per passenger they transport or they just paid for certain roads and a certain time, and so on. So, there are different setups already, and all that will happen in different places and in different ways. We don’t believe in designing a fixed, one solution, but we need to design a solution that is flexible and can be adapted to the specific market.
Ed Bernardon: So, at the other extreme then, I could say, “Provide me all the equipment, train me how to operate it, and I’ll call you if I have a problem.” And there are probably millions of things in between those two extremes.
Felix Andlauer: And even if you would just buy the vehicles and not want anything else, we still think it was a good idea to develop it as a whole because then the vehicle is optimized as a part of the whole and not just to be a cheap vehicle.
Ed Bernardon: The other thing you mentioned was the app. I think that’s really interesting in that I can use your app and I can hire a one-person, four-person. But if I was a city, I might say, “Hey, I want to also use the exact same app to get a scooter or an e-bike or the subway.” Is that a possibility, too, to open up your ordering system or scheduling system so that it could be part of something much bigger for a city, part of the whole mobility ecosystem?
Felix Andlauer: Yeah, and that’s more or less what we believe will happen. So, we will have to have APIs to our service so that someone else can, through those APIs, make a booking of one of our vehicles, for example. And those APIs can be then if a city has one central app that provides you with an e-scooter or with a bus or normal public transport with our service, then, of course, we need to integrate with that. And to be honest, I think that this would be a super attractive solution for us as well because, of course, you lose some of the value chain if you don’t have the interface for the end user. But on the other hand, you don’t have to worry about marketing or anything because someone else takes care of that. So, I wouldn’t say that this isn’t better for us.
Ed Bernardon: And ultimately, it might even include things like an air taxi. So, you have air taxis, scooters, your autonomous vehicle, and public transportation, got all these pieces coming together. Someone might say, “I want the most time-efficient from A to B.” Or someone else might say, “I want the most cost-efficient or the least emissions.” Who knows? And as this thing grows, like you said, in the year 2027, 2028, or 2030, who do you think is going to be the entity, the company, or whatever that figures out these optimal combinations of all these things because not one company is going to provide all this. Sometimes the city provides it, might be four or five companies, the air taxi company, the scooter company, somebody has to optimize all this to give me the most time efficient or cost efficient or emissions efficient. How do you think that’ll play out?
Felix Andlauer: I don’t believe in this one player finding the ultimate solution. I think it’s really like everyone always talks about ecosystems, but then they always want to build it. And I think what’s really the key with ecosystems is that they grow and they optimize themselves. So, I don’t think that there will be one who figures it out altogether, but it’s more like shaping an environment that is constantly changing and where all the solutions compete with each other, and that will ultimately lead to the best solution for everyone.
Ed Bernardon: If you think about what you’re just talking about, it’s so much like what the internet was from initially this thing that “Oh, wow, look at this exciting thing.” And companies and people started to show up on it. And eventually, it just grew, and it governs itself to some extent. They use the platform so that different companies can provide us with information. This is similar in a way, except it’s physical and related to mobility. So, when you hit that critical mass—let’s just call it in the year 2030—it’s almost like you’re creating a physical internet of sorts because that’s what you’re alluding to here. There’s not going to be one person that figures out how to optimize all of this; somehow it’ll grow on its own to do that. What do you think has to happen to make that happen?
Felix Andlauer: It’s out of the question — it will happen. I think the question is more like, how fast will it happen? And I think there’s a lot of these chicken-and-egg problems in the way. For example, developing a vehicle and making it ready for serial production requires a lot of upfront investment. It’s really hard to get this kind of upfront investment for something that you can’t already prove is working. And somehow, you need to have large fleets of serial production vehicles to provide services to convince everyone that this is the solution to our mobility problem. But you need to convince everyone that this is the solution to get the money for the development and so on. Can investors think it’s safe enough to put large amounts of money into it? And will cities or governments provide an environment that de-risks those investments so that they can be sure that it will actually happen? So, there, I think, is a lot of room to change the speed of that transition, but it will happen.
Ed Bernardon: When you are working on the internet and creating a webpage, there is a common language there. It’s the language you use to develop that webpage. Do you think that’s going to happen to the app? We were talking earlier, well, this app controls the NEVS shuttle, but I wanted to also control the scooter. Do you think someone’s going to say—the government or someone—you have to use this type of programming language or this configuration? Do we need something like that, that will make us work more closely together?
Felix Andlauer: I’m not sure. I think, for example, all these kinds of services, the interfaces between them will ultimately be in the cloud anyway. It’s logically already decided what kind of information you will have to send back and forth. So, booking has to include who’s the person you have to pick up? At which place do you pick them up? Where do they want to go? It’s a limited set of information that needs to be transferred. And then if you’re using exactly the same protocol for that or if you have some kind of translator in between that switches around a bit so that they fit, I don’t really think that this will be the big game changer. Of course, at some point, if you work with the same company a lot of times, you will probably agree on a certain interface and use that, and maybe that can become some kind of standard. But I don’t think that this is really the game changer.
Ed Bernardon: I want to ask you about your remote driver system. So, there’s a possibility, at times, where you could take over the remote control of the vehicle. Why do you have that? Because in a way then is that like a backup system in case the autonomous system doesn’t work?
Felix Andlauer: So we did some testing on that just to have a proof of concept. But really, already quite some time ago, we decided that we don’t really believe in remote driving. So we think “remote operator” is the ultimate solution. So, the responsibility has to stay with the vehicle, but whenever there’s a situation that the vehicle doesn’t really know or is not completely sure what to do in — so, for example, there’s a building site somewhere and you have to drive over a line that you’re normally not allowed to drive over, we want the vehicle to ask a human operator, “Okay, I would like to drive here. Is that okay?” And then the operator can acknowledge that or the operator can make a proposal on which route to take. But we don’t want this on a wireless connection to shift around responsibility between the two. So the responsibility needs to stay with the vehicle.
Ed Bernardon: So the vehicle would say, “I need help. Help me.” Rather than a human operator watching all these things and saying, “Hey, I’m jumping in here.”
Felix Andlauer: Yes.
Ed Bernardon: So, the autonomous vehicle has to be smart enough to know it can’t do something.
Felix Andlauer: Exactly. And it needs to get to a safe stop before it does it. So, we don’t think that can be this. The autonomous vehicle is driving around, but we tell this remote operator that if something happens, it’s your fault anyway because you could have done something. We don’t believe in that being safe because that means you always have to have a safe connection, and we don’t see a way to prove that.
Ed Bernardon: So, if in a couple of years, we were to sell a system of 10 shuttles to a city, would you recommend that they have a remote driver?
Felix Andlauer: You definitely have to have a remote operator — someone that the vehicles can ask. And that, I think, you have to have. In Europe, it’s actually this remote operator is called Remote Intervention Operator and it’s part of the autonomous driving permitting process, so you will have to have it. And like in the first pilot, our plan is also to put in a fifth seat in the vehicle and have a safety driver in the vehicle first and start operating the service. Of course, we will lose money on that service because we can’t earn as much money as it costs to pay the driver. But we want to start like that anyway, it’s simply to make sure it’s super safe. And then when we reach a level where we are 100% confident that the autonomous vehicle can handle all situation and is smart enough to stop when it doesn’t know what to do first, then we want to take out the safety driver, but then, of course, we will have an operator. But then probably one operator per 10 vehicles or per 20 vehicles. So it’s very important that you get few operators than vehicles.
Ed Bernardon: Let’s talk a little bit about engineering and designing these vehicles. You mentioned Saab, that a lot of your engineers came from Saab. So that’s going to be your traditional car engineers, ICE engine-type engineers. What’s the percentage right now of traditional Saab engineers versus the new generation of engineers?
Felix Andlauer: I would have to ask HR, but my very subjective guess would be 65% have a Saab history, maybe 70%.
Ed Bernardon: To design an autonomous vehicle, you have mechanical, electrical, software, connections to infrastructure in a city, remote control — how do you get these engineers to work together and collaborate and understand each other’s problems when it’s never really been done before?
Felix Andlauer: Of course, we’ve seen a million problems doing something like that. The car industry is amazingly good at designing cars. But because the whole machine is so optimized to do exactly that, it’s a lot of work to get it to do something else. Of course, we’ve had a lot of problems. We’ve introduced agile management, where the idea is that you have larger cross-functional teams where you get exactly that; people work with other people that they normally don’t work with, so they understand each other, and then you can do something completely new. We’ve overshot a bit and made it a bit too agile for a while, and maybe went too extreme so we moved a bit back. I think we’ve learned a lot, and we are much better at it. This is the kind of work that is never finished. If you find an organization that will tell you, everyone’s happy with our processes and everything’s great, then something is weird about this organization.
Ed Bernardon: They’re probably not telling the truth. Too agile, too much of a good thing. Give me an example of what you saw that was too agile. And for those that might not know what that means, agile development is constantly in contact with the customer, understanding the needs, and showing them prototypes. How can you be too agile?
Felix Andlauer: The hard part is that in developing a card, you have different components with totally different development cycles. So, developing hardware and testing takes months, whereas developing a software feature and testing it can be done in a day. The problem is with these different development speeds, as long as you keep them apart in silos, it isn’t that much of a problem. But as soon as you put them in the same teams, they’re kind of disturbing each other. And of course, if you’re too agile and everyone keeps trying new things all the time, it’s really hard for everyone else — “Okay, which version of your system do I use in my integration?” So, that’s definitely too agile because then you have to balance this continuous development with the crazy idea generation, and you need a little bit of both. And yeah, if it’s too agile, you maybe have too many crazy ideas. And if it’s too rigid and waterfalling, then you don’t have enough ideas. And balancing that, I think, never stops but you have to do it deliberately and think about it.
Ed Bernardon: And waterfalling meaning the order comes from up top. That’s the balance that you have to figure out. When it comes to engineering, what about the engineering processes and the tools that you actually use when you’re trying to bring all these things together? Could you talk a little bit about that and balance that against what you have to do, for instance, to get these people to work together because they don’t understand each other’s worlds?
Felix Andlauer: One thing that we spend a lot of time and money on and that we think you need to do differently is that you need to do a lot of things in the virtual environment. For a traditional car, you know exactly how many kilometers it’s driven and how many times you opened the door, all those kinds of things that you simply know by experience. But in our case, the service we want to design, we don’t have any experience, no one has experience with those kinds of things. So you need to create a virtual environment where you can test your service in a digital twin environment of a city, for example, in order to find out all those things — what’s the right battery capacity, and so on. So, you need to go virtual in many things. And also, when it comes to the hardware design, like in a traditional car, the starting point is much closer to the endpoint because on day one, you already know, “Okay, we want to build an SUV in this segment,” and you have 50 cars to benchmark against, and so you know exactly what the car should look like, and then you’re tweaking a bit the design. Whereas in our case, we’re making much larger steps so we need to have a simulation-driven design. You can’t always design the whole metal structure and then put it into the simulation if it would last in a crash. You have to already let the simulation drive parts of the development in order to quickly get to that optimal solution.
Ed Bernardon: Could you maybe give us an example of the types of engineering tools that you’re using and how they are working together? Because it sounds like it’s not your traditional thing where you just use CAD, say, or you just use an analysis. There are a lot of different types of analysis, there’s software that’s being developed, there is CAD and geometry and all those things. What are some of these pieces? And what have you had to do to make it successful in this mixed environment where you have software, electronics, and mechanical systems?
Felix Andlauer: What we’re aiming for is to have a really digital thread from the city level of a whole fleet operating in a city down to the individual testing of a software component or the individual simulation of one piece of metal. Along all this road to also handle all the time that you put in some requirements in the design, then you design something, and then you have to test against those requirements. And to have these requirements handling enter the digital thread of the development through everything, I wouldn’t pretend that it’s finished because it’s never finished, we keep on working on that. But of course, we try to find partners there that can provide us with, at least, as long as possible part of this digital thread so that the systems naturally fit into each other.
Ed Bernardon: If you were going to give someone some advice that was putting together an engineering organization, what advice would you give them in order to be able to create this digital thread that you talk about in the most efficient way? What are the most important things you have to look at?
Felix Andlauer: I feel like I’m too deep still in that question to be able to give any kind of advice. Maybe you will hear a lot of times that you can’t do something like that because it doesn’t fit the people’s experience. Sometimes the people are right, but I think if you have, as a general rule, to ignore that, then you’re more often right and wrong and just keep on pushing toward this digital integration. Also, when it comes to the toolchain you’re using to go for this minimum viable product approach. So, you don’t aim for a solution that solves everything at once but is there a part of this thread that we can already implement and loop and try using?
Ed Bernardon: It’s interesting what you just said because a lot of times people say, “Oh, this software product, and that solution, and this one.” But maybe more important is to say, “Let’s tackle this part of the problem first, then we’ll add in part B, and then C, and D.” And I would think then that as important as the solution, but probably much more important is to be able to work with the providers of all these different solutions that are capable of connecting things together or capable of helping you figure out how to evolve it. Finding suppliers that can do that and work together I would imagine is important.
Felix Andlauer: Exactly. Even there, it needs to be a partnership because none of the two sides know everything so you need to figure it out together. And there, of course, we’re working closely with our suppliers of these kinds of systems: “We want to use it like this. Does it work? Or do you have any suggestions on what could be done so that we achieve that?”
Ed Bernardon: Let’s talk about the future of autonomous vehicles and where they’re headed. One of the goals you said was you want to take cars off the road. There are about 1.5 billion cars on the road. How do you see that number going down over the next 5, 10, 15, or 20 years?
Felix Andlauer: We normally have this calculation for our kind of service that we think that in a densely populated area, one of our vehicles can replace seven to 10 existing cars; in more rural areas, it’s probably slightly lower. So, what you could say is, of course, it will take a lot of time. But with services like ours, you could probably provide the same number of passenger kilometers with 1/5th to 1/10th of the number of vehicles. However, what makes this whole number game with numbers really difficult is that we have to keep in mind that mobility is not evenly distributed over the world. So, it might be that there will be more vehicles anyway because people travel more. But it could also be that maybe mobility goes down because of the Metaverse, whatever. The absolute number is really difficult. But I think what we could aim for is to reduce the number of vehicles drastically for the same level of mobility.
Ed Bernardon: Like you said, there are a lot of different factors coming into play, all driving to having fewer and fewer cars. One thing you mentioned earlier, though, was Level 4 versus Level 5 — Level 4 being a somewhat constrained area. And if I was the mayor of a city, I would say, “Hey, I want you to work between these two places here.” Do you envision that there’s going to be some sort of a test or a driver’s license? Let’s call it a driver’s license for autonomous cars. Humans have to pass a driver’s license test, does something like that exist now?
Felix Andlauer: In some way, at least the permitting process in Europe already works like that. You have a vehicle acceptance test, where you have to prove that the vehicle as a system can handle all the things, and then you have a site acceptance test where you actually have to prove this vehicle can handle all situations that can occur in this area. And there, it’s so important that that’s why this area will have to be limited and defined in the beginning because proving that you can handle all situations that can reasonably occur in an unlimited area is an almost unlimited problem. But as soon as you can reduce that to a smaller area, it’s pretty much what you do. You have to show, in simulation and in real-world driving, that all situations that happen in this area, the vehicle can handle. And if you think about it, a driver’s license is not that much of a difference. If you take your driver’s license test, you drive around and you have to prove to the tester that all those situations I can handle.
Ed Bernardon: We’ll expect the same from autonomous vehicles as we do the human. So if they have to do a driver’s test, then the autonomous car has to do it as well. And certainly, the physical test makes a lot of sense. If I’m driving from A to B and there are stoplights or there are traffic lights, or whatever, you could set up some sort of test environment. Do you foresee, also, the potential of some sort of a standardized test in a virtual environment? We talk a lot about digital twins that say, “This is the test that you must pass.” And maybe you plug in certain things into that environment, like the types of streets or the number of people or the other types of vehicles. Does such a thing exist? Or do you see something like this, if it doesn’t exist, having to evolve over time?
Felix Andlauer: Again, it’s not my area of expertise, but to some extent, it already exists but it’s not actually describing the exact situations you have to handle but it’s about the standardization is around the process — the process how to describe what can occur in this defined area. For example, if in the area, there’s a roundabout, then you have to prove that you can handle a roundabout; if there are traffic lights, you have to be able to handle that. But if there are things that you can’t handle but you can make sure they don’t exist in this area, then that’s okay as well. So, I don’t think there will be a standardized test where exact scenarios you have to drive because then it will be super easy for any AV company to optimize for that test but it won’t tell you anything about how the vehicle actually behaves.
Ed Bernardon: I suppose you had to have a standardized methodology that says, in order to evaluate the items or sometimes people use the word “actors” or “scenarios” that you have in your test, you somehow have a process there where you look at the physical environment and say, “Oh, this physical environment requires these things in the test, let’s go do it.” But that does not exist yet, or does it?
Felix Andlauer: There’s at least a lot happening when it comes to legislation in Europe. So I would argue that, of course, the exact details you’re still kind of figuring out while you’re doing it, but the legal framework is close to what you’re describing, I would say, at least in Europe. We see a lot of autonomous vehicles in the US because the legal logic is kind of different in the US — very simplified, it’s more like everything is allowed until someone proves it should be forbidden. Whereas in Europe, everything’s forbidden until you prove that it needs to be allowed. And that leads to a higher threshold in Europe to get out. But when you’re out, it’s actually, from a commercial point of view, much less risky because you only have to prove that “Okay, everything I’ve promised, I can do. But if something happens that I haven’t promised and no one has thought of, it’s not that much of a legal problem like in the US.”
Ed Bernardon: So, as a businessman, do you prefer the US approach to that or the European approach?
Felix Andlauer: It kind of connects to, also, the investor environment. I think European investors wouldn’t invest in the American environment because it’s so risky. Being a European, I think I would invest my money in a European AV company and not an American.
Ed Bernardon: Less risk. All right, it’s the year 2050, way into the future. Where are you located in Sweden?
Felix Andlauer: Our headquarter is in Trollhättan, on the west coast, but I currently actually sit in our office in Gothenburg.
Ed Bernardon: What do the streets of Gothenburg look like in the year 2050?
Felix Andlauer: I don’t have the slightest idea, to be honest. I can forecast for 10 years, I would say all the vehicles out there are more or less autonomous vehicles. You will hardly see a person driving a car anymore. But beyond that, stuff like the metaverse, people meeting in virtual space instead of physical space could do so much to transportation. Maybe no one wants to live in a city anymore because it’s become too hot and everyone wants to live in the forest.
Ed Bernardon: We all live in the metaverse. Felix, thank you so much. It’s been a great conversation. It certainly opened up our eyes and our minds to some new ways of looking at autonomous cars.
Felix Andlauer: Yeah, that was great fun. Thank you very much.
Ed Bernardon: So, let’s finish off, as always, with Rapid Fire — some quick questions with quick answers. What was the first car you ever bought or owned?
Felix Andlauer: It was a red Nissan Mitra. I actually owned it together with my brother, and my brother put large flames on both sides, so I got stopped by the police a lot.
Ed Bernardon: Do you have large flames on your car now? Or do you plan to put those on the shuttle? That would be a real feature.
Felix Andlauer: I will talk to creative design about it. I’m not sure if I can get this through.
Ed Bernardon: Maybe you could have a reconfigurable skin so I could order it with flames when it comes to pick me up. Did you pass your driver’s test on the first try?
Felix Andlauer: Yes, of course. I’m a smaller brother and I’m super competitive, so no way I would miss a test like this. No way.
Ed Bernardon: Have you ever gotten a speeding ticket?
Felix Andlauer: Yes. Not many because all my grown-up life, or most of it, I’ve lived in cities, I didn’t own a car. So I don’t have many, but a few of them I have, yes.
Ed Bernardon: Do you mind telling us your best speeding ticket story?
Felix Andlauer: It’s maybe not great. One time I remember that I talked to a friend that I wanted to visit while driving to him, and I told him approximately where I was, and he told me, “Oh, be careful, there’s a speeding camera there.” And exactly the moment he said that, it flashed. So I have a picture of me holding the phone while him telling me that I should be careful.
Ed Bernardon: In three words, how would you describe what the vehicles of the future will look like?
Felix Andlauer: This is really tough for an engineer to answer because I really don’t have a good vocabulary to describe design. So my first word would be “reduced.” Think of an iPhone, it doesn’t look fancy. Most phones look more or less the same because they want it to be very clear and maybe Scandinavian, I don’t know. And the second one would be “boxy” because I think in the service world, everything will be about ingress, egress, and comfort. So, these super low sports cars that you more or less lay in, no one will want that. The third one is “comfortable.” When I talk to creative design and I ask them, “What does the latest design look like?” I want them to show me the interior and not the exterior of the vehicle because I think that’s what will matter. And again, the inside will look super comfortable because that’s the differentiator.
Ed Bernardon: Smaller, boxy, and comfortable. So, we always ask people on The Future Car podcast, what their living room on wheels is going to look like. And this time, we have an expert in autonomous cars with reconfigurable interiors. So if you had an autonomous car that you could take, in this case, let’s say it’s on your own, a five-hour trip, what would you have inside your living-room-on-wheels autonomous car?
Felix Andlauer: It’s really funny because we’ve built this first prototype of our Sango vehicle that we call Release Candidate 1, and that’s pretty much what I would want it to look like. The vehicle doesn’t have any screens. All the screens that are there are below textiles. So, if there is no information to show, you’re not looking at a blank screen, and it’s just super calm and relaxing. If you ever come over to Trollhättan, I can show you my dream living room on wheels.
Ed Bernardon: You’re probably our first guest that actually has one. What person, a person that’s currently alive or someone from the past, would you want on that five-hour trip with you?
Felix Andlauer: I would go for Barack Obama. I think he has a lot of interesting stuff, really top-secret gossip to tell, and I think he’s also good at being silent. Because if you’re on a five-hour trip, sometimes you just want to look out the window and don’t say anything, and this would be really important criterion for the person I choose.
Ed Bernardon: Maybe he could give you some tips if you want to get into politics, too.
Felix Andlauer: I’m pretty sure he could.
Ed Bernardon: If you could magically uninvent one thing what would it be?
Felix Andlauer: It’s really difficult because, for me, I totally believe that technology normally makes people’s lives better. Of course, there are always extremes and edge cases. But if there’s something I could unmake, it’s not really an invention but I would go for social media. That’s one of the few things that actually do more bad than good.
Ed Bernardon: Yeah, we get that answer a lot: social media. So, what would you invent? If you could magically invent one thing, what would it be?
Felix Andlauer: A lot of these things are already happening, but probably some really cheap, renewable energy source that solves all energy problems at once.
Ed Bernardon: Last question: Tell us something about yourself that would surprise your friends and family, something they don’t know about you and now they will after you tell us.
Felix Andlauer: I’m talking so much and I’m just telling everyone everything anyway. I wouldn’t expect anyone who knows me well that there’s anything that they don’t know about me. I think the best I can come up with is probably my family would be surprised at how I behave at work, and my colleagues would be surprised at how I am at home because at work, I hate procrastination, I am very orderly and everything needs to be structured and so on. And at home, I’m the complete opposite.
Ed Bernardon: You’re a big procrastinator?
Felix Andlauer: At home, I am, yes.
Ed Bernardon: What’s the biggest thing you procrastinated on in the last week?
Felix Andlauer: I’ve lost my wedding ring in the car below the driver’s seat and his has been sitting there for four weeks. I can’t just get it out, I have to unscrew the driver’s seat but I haven’t managed to do that in four weeks.
Ed Bernardon: There you go. Felix, thank you so much for joining us on The Future Car podcast.
Felix Andlauer: Thank you so much.
Felix Andlauer, VP Mobility Solutions at NEVS
Felix Andlauer studied mechanical engineering in Munich and Gothenburg. From 2007 he worked with renewable energy in a variety of positions combining technical, business development and leadership tasks. In 2018 he joined NEVS as project manager for the newly started mobility ecosystem development project PONS. Since May 2021 he has been VP Mobility Solutions at NEVS.
Ed Bernardon, Vice President Strategic Automotive Intiatives – Host
Ed is currently VP Strategic Automotive Initiatives at Siemens Digital Industries Software. Responsibilities include strategic planning in areas of design of autonomous/connected vehicles, lightweight automotive structures and interiors. He is also responsible for Future Car thought leadership including hosting the Future Car Podcast and development of cross divisional projects. Previously a founding member of VISTAGY that developed light-weight structure and automotive interior design software acquired by Siemens in 2011. Ed holds an M.S.M.E. from MIT, B.S.M.E. from Purdue, and MBA from Butler.
If you like this Podcast, you might also like:
- Bringing Sustainability and Privacy to Autonomous Ride-Sharing with Felix Andlauer – Part 1
- Carlo Mondavi’s Autonomous Electric Tractors for Sustainable, Affordable Farming – Part 1
- The Next Leap for Electric Vehicles with Will Graylin, Indigo Technologies – Part 1
The Future Car Podcast
The tech-driven disruption of the auto industry cuts across domains, from silicon and software to sensors and AI to smart traffic management and mobility services. Get the chip- to city-scale story in regular interviews with technologists at Siemens and beyond.