In the final episode of the Future Car series on the transportation revolution, Nand Kochhar and I take a closer look at autonomous vehicle development around the automotive industry. In this episode, we learn about the dynamics of autonomous vehicle development, and how they may fit into transportation systems in the future. Nand also provides deeper insights on AV verification, testing, validation, and how confidence in this new technology is created among consumers and companies. The transcript of our conversation follows:
Conor Peick: Alright. So, Nand, start off our discussion of autonomous vehicles and autonomous drive, of course. Could you just give an overview of the different levels of autonomy and where we’re at on the spectrum right now?
Nand Kochhar: Yeah, sure. As you know, there’s five levels of vehicle autonomy in terms of the SAE definition – level zero is no automation, but level one is what we call a driver assistant system. The level two is partial automation in which you can do, let’s say, the steering and the braking functions together. By now, the majority of the OEMs have matured to that point that they are offering SAE level two. So, the examples from a usage standpoint would be lane-centering or lane departure warning, doing some steering functions related to that. The example in the braking would be autonomous emergency braking or the adaptive cruise control, which most vehicles have today. To my experience, that SAE level of autonomous maturity is out there. Now, next comes the SAE level three, which is the conditional automation scenario, where you also have the environmental detection capabilities. In other words, you have the mapping function and then you can take a car or a vehicle from point A to point B, and now you’ve matured it to level three, but you still need a human override, so a lot of the companies are offering different versions, different levels. Level three is the most confusing because there’s so many scenarios that they allow you to operate but they want hands on the wheel all the time, or they’re doing a face recognition and looking at the driver, so the driver needs to be attentive; you can’t go to sleep or reading a book as an example – the level three, what’s called conditional automation. So, we are in that journey right now. Global companies transitioning into level three, they’re claiming that these are the level three vehicles and they’ve got certain scenarios up to some extent, offering those. The next level is level four, which is high automation, where you can truly do these functions I talked about but still a human override is still an option. So, in other words, a vehicle is confused about something, it’ll give the warnings and then human intervention is required; that’s level four. And then level five becomes the ultimate where you have no constraints, there’s no geofencing, there are no conditions. You can work in all operating conditions, whether it’s day, night, etc. That’s the full automation, SAE level five, which is in my mind, far out, further out, than any of the other levels. So, those are the high-level definitions of SAE level, zero to five.
Conor Peick: Maybe just as an example, I’m sure pretty much everyone is familiar with Tesla’s autopilot system, and there are competitors coming out for that as well, but where would that fit on the scale?
Nand Kochhar: That would be in that level three category I talked about. Again, every OEM might have different conditions for other reasons. For legal as well as for what’s required by law in different countries. What the driver scenario is, the warning systems are, whether they allow him to keep hands on the wheel all the time; I have done some vehicles where you can take your hands off for a few seconds and then you get a warning that puts your hands back on the vehicle.
Conor Peick: So, you just mentioned when you were describing the level five, you talked about different weather conditions, and then, of course, driving in the day versus the night. I know most people understand this is quite a different task. That kind of brings up, I think, one of the most difficult parts of validating an autonomous vehicle is just the sheer number of different scenarios that one might encounter out on public roads. So then, how can a passenger, or for that matter, a company be confident that the vehicle will be able to perform safely and comfortably at all times, given these just the multitude of different scenarios it might encounter.
Nand Kochhar: Yeah and you’re definitely right on that it’s a daunting task of all the possible scenarios. No matter how many scenarios you imagine, there could be still some where we call the corner cases you might find that you haven’t thought about or the systems haven’t thought about. That’s where I think the simulation technology comes into play, that instead of relying on purely only physical testing of all these scenarios that might not even be possible. Over the years, there have been several studies that would say, how many years it’ll take and how many millions of miles you have to do a validation testing if it was a physical-only that companies are adapting, and the technology companies including Siemens is offering solutions in the simulation world where you can create these scenarios and create what’s called the edge cases, the things we haven’t thought about. That’s where the other technologies like machine learning and artificial intelligence are coming into play to create those scenarios. So, in that case, you run your vehicle during the development cycle through those scenarios in a virtual environment. That’s the best way you can do in terms of assuring that you accounted for maximum number of scenarios possible in that given condition. So, simulation plays a big role in that autonomous to the extent that I can pretty much say that having done simulation for over 30 years, that it’s not even possible without simulation when It comes to the autonomous or ADAS vehicle development.
Conor Peick: Interesting. Is anyone taking a purely simulation approach, or usually it’s a hybrid kind of between simulation and physical testing?
Nand Kochhar: The big challenge in simulation is that at the end of the day, people don’t buy virtual models and they don’t drive. They drive real products, whether it’s a truck or cars or whatever. So, I think the key thing is the correlation of a physical test and the virtual test, and if they’re correlated, that’s where you build your confidence. In Siemens terms, we call them digital twin, and the industry also is getting very popular on digital twin, but what we do is the most comprehensive digital twin, a performance, that what you get a performance level, which is physics-based in a virtual environment, you can get the same performance of the physical test. Once they are correlated, that’s how you build confidence, and then you can start relying more and more on the simulation or the virtual testing. And then there were several other scenarios in between, that it’s always a hybrid in the sense that you can bring a hardware in the loop, whether it’s starting from scratch, developing a chip, or a system on a chip, to a full vehicle, and you can have a vehicle in the loop systems as well. So, you’re right. It’s a hybrid because of getting that correlation and building the confidence.
Conor Peick: How do you see that balance between the physical and the virtual testing? How do you see the balance evolving moving forward? Is it leaning further into the virtual testing or is it staying pretty even on either side?
Nand Kochhar: I don’t know. I think it is growing quite a bit in the virtual environment because in addition, too, some of the scenarios are not even possible and the physical testing environment. Let’s say if you want to do a car stopping when a pedestrian is crossing the road, in a physical test environment, to have that setup, it is a tremendous amount of investments in the prototypes, as well as in the production systems as an example. In the virtual environment, you got initial investment, but after that, you can run all these iterations in a very very economical way, that it’s real and you can use that for production. So there’s more opportunities, more growth going on in the virtual environment, in the simulation environment, in my mind, but of course, at the same time, the physical side of the business is maturing so that they can keep up to date, what’s being done in the virtual environment and then the correlation happens, because at the end of the day, certification is still a physical testing based for most parts of the world. So, we do need to do the validation and certification as a final run. So, I think one of the ways you do that is you want to make sure your first time through physical testing is a success, and you run all these scenarios in a virtual environment to play those scenarios so you have a very high success rate, and a physical test environment is a good way to balance that.
Conor Peick: So, you’re using the virtual to prepare yourself for the physical test.
Nand Kochhar: That’s right.
Conor Peick: Stepping back somewhat from your level five autonomy. Today already, there’s a large number of assistance systems available, you mentioned the lane-keeping assistance and automated emergency braking. And as we see, a greater number of these systems become available, we’ve also seen a lot of additional regulations come out. So then, I suppose, what do you see is the effect of these new regulations on the future of vehicle validation and certification?
Nand Kochhar: So, the future regulations have to take into account the maturing technologies. Today, for example, in safety, there’s a well-established – and I’ll use the US example, again – NHTSA National Highway Traffic Safety Administration, what the protocol is for those testing. In addition to that, they also have the rating system where they give vehicles one to five-star ratings from a passive safety performance. That’s where your airbags, seatbelts, performance, etc., comes into play. The regulations are starting to mature; they’ve started to take into account the active safety features and how they play in the car. Those contribute towards the safety ratings, is a good example, that regulations are adapting and maturing. But to the best of my knowledge, today, there is no test procedures declared which will be accepted for a pure autonomous vehicle running. So, there’s a lot of work to be done. In addition to the technology side, both on the regulation and the legislation front, policymakers, a lot of work is still needed to be done in that arena.
Conor Peick: In the development of these systems, these ADAS systems, how can companies find a balance between proving safety and compliance, and meeting these regulatory requirements, and also satisfying the business challenges, wanting to produce a high-quality profitable product?
Nand Kochhar: So, there’s parts of it. If you look what are the key components of making ADAS features and autonomous vehicles, we can have an analogy of a human body – there’s the eyes or the perceiving side of it, which is the camera, lidars, radars; they are bringing in the signals through the sensors. And then you have a processing of that information, which is the equivalent of brains for a human driver in the car, and for that, you have several algorithms and the computing environment in the vehicle, or it could be on the cloud, to make those decisions, what to do with those scenarios? And the third part is the commanding as a based on that scenario, what do you want a vehicle to do? That’s the muscles in a human body. So, you want the brakes to apply? Or you want to steer the vehicle? Those are the examples. I think we continue to make progress in each one of these as the volumes grow for, let’s say, the camera, lidar, radar technologies maturation, cost keeps coming down. You will see at some point that it’s a balance that customers are willing to pay for new technologies initially, and then as it matures, you get more and more higher volumes, and the cost keeps coming down.
Conor Peick: Do you think that fully sort of virtual certification of ADAS, or for that matter, maybe even an autonomous vehicle someday, do you think it will ever be possible to certify these technologies virtually without any physical testing?
Nand Kochhar: Yeah, being the simulation kind of guy, I think it is always our vision to mature the simulation technology to that level. You might take a while to get to the SAE level five, that’s where the levels come into play, but on some of the other feature functions, over a period of time, you will see that some of that would be possible in a shorter period of time.
Conor Peick: I’m sure the company building these technologies would be happy to hear that outlook because, as you mentioned, it’s much more cost-effective than doing physical testing.
Nand Kochhar: That’s right. The road is not a glorious road, there’s a lot of investment in doing that physical test correlation, taking care of all the noise factors or the other variabilities, and building that confidence. In this case, would be also bringing the regulators along with you during that development journey. So, they are part of it, and so they can build confidence that what’s being done in a virtual environment or a simulation environment. They can go through the assumptions, which are being made. They can do the testing around the confidence level at each stage or each level, and then get to that point where you can do the certification.
Conor Peick: When it comes to the simulation, maybe the simulation engines, you mentioned two aspects being machine learning and artificial intelligence and then also the fact that they’re physics-based simulation. I was hoping you could maybe describe how each of those aspects kind of feeds into building these more and more sophisticated simulations.
Nand Kochhar: So, with the advances in computing power, I think that is what is allowing all the other technologies from an application standpoint. The physics-based model is that when you have a sense of model on your vehicle, it is not simply reflecting from the city infrastructure, let’s say, a computer gaming model versus you have a model of the real city, and the properties of the building, of the roads, or the trees built in those models. So, that’s what is a physics base. When there’s a signal going hitting that building when it’s reflecting and coming back, it is not just an imaginary graphics purpose, a gaming engine, it is more of a, “What’s really happening in that environment?” You can do that competition with them as you need to build a huge model of the city infrastructure, as well as all of the different materials, the different possibilities of hitting these at different times of the day – where the sunlight? What the reflection scenario is? – So, you can imagine, that’s huge. In terms of machine learning, and artificial intelligence comes into play in a lot of those, building up those scenarios. So when you’re driving around what we touched on previously, what we call the corner case scenarios, you could have sensors on a vehicle while running on a road, collecting all this information. They’re feeding that information back to the cloud infrastructure, and then comparing that into what the baseline is in. With that, they’re doing self-learning and recording, and they keep getting better as they come across other scenarios. So, that’d be a good example of how the machine learning plays in the self-learning and adapting to the different scenarios, and then using that information for the next possible development in the vehicle environment.
Conor Peick: So, you’ve talked about comparing between the simulation and the physical environment and how using those comparisons or the correlation between the two, you can continue to build up the accuracy, or at least your confidence, in the simulation. Does machine learning help accelerate that process?
Nand Kochhar: Yes, machines learn those several scenarios you can create. Machine learning is, in a simplistic way, nothing but you observe the previous scenarios, and instead of just the human looking at it, the machine learns in terms of zeros and ones of that information for a given input, what the outputs were, making it very simplistic. You create enough scenarios or enough iterations that your database is strong enough to handle the next set of scenarios. So, the next scenario you feed in has never happened in a physical environment; it’s just a virtual feed. And based on the previous learnings, the software will tell you what the outcome would be. That’s where you build the confidence that now you can rely on that because you’ve had enough learning by feeding in all the previous information. That is one example of machine learning.
Conor Peick: So, the next, maybe we can kind of pull back out to the industry-level view. One thing that’s been notable with the development of autonomous vehicles is the emergence of some potentially unusual or unexpected partnerships. I think one notable one was that Mercedes and BMW announced they would be co-developing some autonomous technology. So, what do you think is the impetus between these partnerships between what you would otherwise think of as fierce competitors?
Nand Kochhar: Yeah, there’s a couple of things. The challenges are so huge, and there’s a, so to say, skillset shortage, and everyone feels a need to get on and make some progress. And then also, it’s expensive even when you do have skill sets available. It’s very expensive to build the teams with those skills and the infrastructure needed, whether it’s the computing infrastructure or the testing infrastructure, etc., that’s huge. So, companies are forming all kinds of partnerships so that collectively they can leverage what’s, so to say, common from a technology standpoint doesn’t interface with their competitiveness, from the IP standpoint, intellectual property standpoint, and they build these partnerships with the hope that they can, both of them in these partnerships can progress. We used one example, and you’ll see several of them with the technology companies and an OEM, between two OEMs example we used, and one good one is Volkswagen-Ford working with Argo AI, doing a common investment; that’s how they’re going to mature their autonomous technologies. Hyundai working with Aptiv, that’s another good way of saying two different aspects coming together to deliver the autonomous vehicle technologies. So, you’ll see several of these but the common driver is speed at which both partners want to progress, and also they don’t want to be left behind on delivering the autonomous vehicles. So, those are some of the reasons behind. Obviously, at the end of the day, it all has to make a good business sense to make these partnerships successful.
Conor Peick: You mentioned that the skills gap or the lack of the necessary skills in the sort of automotive industry, do you think this is an area where other technology companies can… it gives them a bit of an advantage, perhaps, as they try to get into the vehicle space?
Nand Kochhar: Definitely, I think that technology companies have done a huge amount of transformation within whatever part of the business they are in, whether it’s software or whether it’s artificial intelligence, it could be in telecommunications, it could be in purely the Facebooks and the Googles of the world. So, they have a very strong foundation and technology, and it’s all software-driven as well. So, what’s common in vehicles are becoming very software-centric automobile. You could say it is becoming a software on wheels, to some extent, or it’s your IT infrastructure on wheels because there’s so much computing power, and so much software available. In an autonomous vehicle example, you will have software, what’s called the embedded system software, within the vehicle. You’ll have software for infotainment, and you’ll have a software with connectivity even to do your shopping scenarios within the vehicle because you’ll have a vehicle talking to Amazon as an example. So, as you can see, it is a common element. The second part is that technology companies see that the automotive industry, especially the way it is shaping up as connected, and a business on the wheel is a predominant industry and is a good business case. So, they’re bringing not only the technology but their money as well, to invest in and for the future. I see a huge advantage technology companies have, and that becomes a big challenge for traditional OEMs because some of the technology companies are not coming up with any legacy issues, whether it could be labor-related, whether it could be the current employees related, etc. They are, so to say, relatively fresh. They’re 20-30-year-old companies versus 100-plus-year-old companies. So, they bring in a lot of challenges to the traditional OEMs because they have a new way of thinking, out of box thinking, a new way of doing things and they have experience in the technology part of it, which is these two fields, automobile and technologies, almost all come together.
Conor Peick: Earlier we were talking about… or maybe I think it was perhaps the last episode we were talking about over-the-air updates and the potential availability of add-ons, almost, where a customer could purchase additional capabilities or features for their vehicle. This is an area where if you think about a mobile phone example, you have Apple with the App Store, that’s maybe something that they have experienced maintaining that could be a huge advantage in developing some of these new automotive business models.
Nand Kochhar: Definitely you used a very good example. So, how many times when you get a new iPhone, you had to take it to a local dealer to get something done? No, it’s so user friendly; these things happen behind the scene. Most of the time updates are happening while you’re sleeping late at night. So, to them, it is very natural because they’ve been doing it for years. For automobile’s over-the-air update, very few companies are doing today, and every time it’s big news. I’m sure, when we touched on Tesla, that’s a big differentiator, but obviously, others are catching up. But it’s still a big deal. It’s not that easy as your iPhone getting updated and nobody even notices.
Conor Peick: Right. Because, of course, an update on a vehicle, you have to make sure that it hasn’t somehow impacted any of the other may be more critical software like one that runs the emergency braking system, for example.
Nand Kochhar: Absolutely. So, the safety vehicle is different than your iPhone, you could use that analogy that it’s not the same as updating an iPhone because there’s a safety factor and you’re driving and you’ve got vehicles in there with the people, etc. But when you look at the pure technology angle, then you say, “Oh, that’s a software update.” But you just need to obviously put extra security, extra checks, and making sure it’s not interfering with the systems. But from a technology standpoint, there are a lot of similarities.
Conor Peick: So then moving on. I think, certainly, everyone recognizes that level five autonomy is still ways off in the future but is, of course, the end goal for a lot of these companies. What can we do in the current environment with the technology available today what can we do to push the boundaries of autonomy and continue to develop the technology?
Nand Kochhar: Yeah, so the progression to these levels, that’s one approach that you go from level two to level three. I’m not even talking about level one to two because there’s a settle that’s been done, so to say, for a good chunk of the companies, that you continue to mature, but there are four things in my mind and technology is just one piece of it, that technology needs to continue to grow, enhance, making sure of the security and the safety is paramount, and you can start addressing the maturing of that technology. The second important point is about the business model and the business case, that when we are doing these things, making sure autonomous still delivers profits for the companies and also delivers value for the users or the consumer. The third and the fourth factor of bringing in a shift and a change of the level of autonomy of level four or level five comes into account is the legislation and the policymakers. Those are big deals in my mind, if you put as a percentage of each factor because at the end of the day if something does go bad, who’s responsible for it? Who’s accountable for it? What are the legal implications? All those things make a big, big factor in terms of the policies, and the safety, and the security, etc., are the factors which will drive how we speed up this from level three to level four, and someday from level four to level five. Also, from a pure technology perspective, there might be a shift of technologies as well. The current technology approach of growing one by one level, there might be newer set of technologies in the works which will jump directly to level five vehicles and might offer a different set of technologies to be used. I think some companies will come up with those ideas as well, and that’s yet to be seen, how practical will that become. So, those are some of the factors in there. But that is a holistic picture, not just purely a technology or a user business case, and combining that with the policymakers, combining that with the certification homologation, keeping that in mind. This is where you see a little bit of the variation when you take a global look. Some countries will take that leap, and they will adapt these levels of autonomy sooner than the others. Some will take chances because all these other policies and legalities are different in different countries in different parts of the world, and that’s where you see another big variation, that you always have to have that global viewpoint in mind, not only a holistic focused only on one area.
Conor Peick: So, it’s not just a question of technological advancement. It’s also a question of society being ready, I suppose, to accept and use these vehicles.
Nand Kochhar: Yes, it’s the society being ready and everyone else supporting the society. What role they’re playing, for example, education is one of the big things. When you go take a ride in an autonomous shuttle, which you can do that in Las Vegas here, there are a few shuttles running around. When you experience that for the first time, you have a different opinion versus once you’ve done it 10 or 15 times. So, that becomes part of the education. Society, in general, I’ve come across a few interesting scenarios, how is normal population behavior in general, and in other words, you will see autonomous vehicles will have a flashing light, telling pedestrians it’s an autonomous vehicle, and people will pay attention to that. It was an interesting scenario shared in certain parts of the world, people will see that there’s the autonomous vehicle and they want to jump in front of it just trying out, would it really stop or not? Now, you and I will never think about that in the US, “Let’s try it out.” Now, that’s the education part. That’s in certain parts of the world, people are more gutsy, they say, “Yeah, I understand. Now, let me see if it works or not.” They don’t worry about the consequences, just in case it didn’t work, what would happen? So, these are little interesting examples you see that there are so many other factors, not just the technology.
Conor Peick: That seems like a very high stakes gamble to take for very little payoff.
Nand Kochhar: Yup!
Conor Peick: So, I suppose, getting back to the technology side, however. one aspect or I suppose a potential approach that we’ve heard about is the idea of remote monitoring of the vehicles, that being a way that could bring us closer to level four and level five. What do you think is the value in that approach and what are the challenges involved?
Nand Kochhar: Yeah, definitely. So, these are all scenarios of geofencing. How do you go from level three to level four? Geofencing is that you define an environment in the remote monitoring, you could have one of the environments. So, the advantages you’ll have in there is the ultimate goal – one of the goals is you don’t have to have a driver in the vehicle, so they’ll allow you to do that. So, that’s a big plus. On the other hand, if there are situations, which a vehicle is having some difficulties by remote monitoring, you can address those and sometimes even control those vehicles remotely. So, it gives you a safety net, so to say, in that environment. So, now you’re pushing the technology into an autonomous level. You’re getting advantages of what one of the goals is, and then still making sure there’s a safety net before you let it loose completely. That’s where I think the legal part might come into play and the policy by different countries. Some countries might allow that, “Okay, you don’t need the remote driver after you’ve been driving for X number of months or years.” Other countries might say, “No, I always want you to have that backup.”
Conor Peick: That also brings up the aspect of security. This is probably even true with connected cars today of, you know, how do you ensure that the remote monitoring system or the over-the-air update system are secure and not vulnerable to outside actors?
Nand Kochhar: Definitely a huge issue in terms of making sure it’s addressed before anything gets, so to say, practical in terms of the mass distribution of these types of vehicles. The safety security starts to come into, from the very basics of software development for these vehicles. There are new compliance methods and standards being put in place, for example, for software ISO-262 compatibility to assure the software is going to work and to some extent of security as well, and then a lot of the regulations and procedures around cybersecurity, they play a big role and not only at the end state of application and usage but also during the entire development process. Making sure on all steps along the way that thing is taken into account.
Conor Peick: Do you think that geofencing and remote monitoring, is that a plausible way of maybe kind of lowering the barrier to entry, I suppose, for autonomous vehicle deployments, making them easier to get achieved?
Nand Kochhar: Yeah, definitely, it is a way of achieving that and also making people comfortable with it. You talked about society and providing advantages. Several cities have come up with dedicated lanes for electric vehicles, autonomous vehicles kind of thing. That could be one of the ways of making sure that people are starting to get comfortable, and they see some advantages as well. It’s just like in LA, your carpool lanes, that kind of set up.
Conor Peick: So then, speaking of this society perspective, again, how can we push for the development of regulations and legislation, it’s an aspect that we’ve mentioned a couple of times has been really critical to the actual adoption of EV. So, are there practical steps that can be taken to improve, or maybe not improve but enact policies around autonomous vehicles?
Nand Kochhar: Yeah, I think collaborating and working with the standards organization, for example, again, SAE International – where I’ve spent quite a bit of time in leadership roles – plays a very important role in setting up those standards and implementation of those and then working with the government agencies through SAE could be one good example of how these things could be continued to be enhanced and bring to reality. I’m sure there’s other parts of the world or other standards, organizations like ISO. And then in Asia Pacific, there might be the local authorities having their own standards organizations. That’s one way of approaching, getting the legislation building bodies, letting the policymakers engage through those processes, and then making progress on that.
Conor Peick: We’ve also talked, of course, about technological development. The other aspect to that is, as you’ve mentioned several times, the business development. How are maybe business challenges either accelerating or holding back these emerging trends of autonomy and self-driving vehicles?
Nand Kochhar: So, I think business trend is always central to making any technology available for masses. At the end of the day, today people are willing to pay for that technology, but that’s servicing a much smaller percentage who can afford to pay higher pricing. You used the example of Tesla before, and it is a little bit more expensive than the traditional vehicles but people are willing to pay for it. Now what you want is to grow that number of people or the volumes so that these technologies become available for the masses. That happens in two ways. One is to grow the economies of scale – someone needs to continue to invest in the technologies and mature them to a point that you have both from a manufacturing standpoint and a production standpoint you grow the economies and the cost curve keeps coming down. That’s one big factor which will bring these to fruition.
Conor Peick: The other aspect is the development of city infrastructure for helping autonomous vehicles navigate within dense urban environments. So, how the development of those technologies plays into this?
Nand Kochhar: Of course, we touched on not only from an autonomous perspective but also the connectedness and in multimodal transportation, all those become part of the connecting to the cities and the city infrastructure. It becomes a very important part for the autonomous vehicles. So, if you’re pulling into a restaurant and your vehicle is fully autonomous or even a level four autonomous that the self-parking or some of the companies are offering the feature called valet parking; I think, what a convenience to have. It’s pouring hard or it’s snowing and you’re done with dinner with your date, and all you do is push a button on your cell phone and when you walk out vehicle is sitting there waiting for you all warmed up heated, and there you go, you pull in and go. You don’t have to turn your ticket in, to the valet guy. You don’t have to worry about waiting there in the cold, etc. So, you see how the city infrastructure, in this case, could be the parking structure nearby. The restaurant owner and the customer going into that, into an autonomous vehicle, having the network connectivity, whether it’s 4G or 5G there that the mobile phone is able, is confident and it will be collected all the time. Now you got the entire chain built. And then you don’t have to worry about driving late at night in the dark either because your vehicle is autonomous, and it’s going to bring you home safe.
Conor Peick: Are there good examples of companies that are leveraging these infrastructure technologies in deployments of autonomous vehicles, even experimentally?
Nand Kochhar: Yeah, you will laugh again. Tesla is doing, I think, valet parking. They are bringing in and people are starting to use it at a consumer level. Some other companies I’ve seen from a research standpoint have been doing it for years. The valet parking example that you stand by your door and you push a button and your car comes to you. So, now it’s just a matter of maturing this to production level technologies.
Conor Peick: So, then, Nand, the other major automotive and transportation trend often mentioned last in that kind of list of different trends is the idea of shared mobility or mobility as a service. How critical are autonomous vehicles to this concept? Or is shared mobility or mobility as a service is something that would bring autonomous vehicles into the mainstream?
Nand Kochhar: So, mobility as a service, there are several aspects, but obviously, autonomous vehicle plays a big role in some of those scenarios. A good example to use is a shuttle service within a city running all the time, that would be autonomous. The shuttle of, let’s say, at airport terminals running from one end to the other end is shared vehicle, there’s multiple people in it. So, those are some other good examples of shared mobility and mobility as a service. Autonomous plays a big role. You can also see the rental car model changing – you rent a car, and if it’s autonomous driven, it gets delivered to your home, and you take it where you need to go, and you send the car back when you’re done with it. So, those are all very good practical examples of mobility as a service and shared mobility. In addition to all the delivery capabilities, whether it’s delivery of an Amazon delivery truck, or a drone, or delivery of a pizza from a local pizza store. Those are some of the things some other companies have already experimented. So, I think, yeah, a huge topic on its own, and love to talk about it our next time.