HxGN RADIOPodcast

Enabling autonomous vehicles with safety-critical algorithms

Apex.AI’s innovations extend autonomous vehicle technology’s potential to benefit a variety of vertical markets such as automotive, farming, mining, construction, aerospace, transportation, manufacturing and more.

BK: Apex.AI’s innovations extend autonomous vehicle technology’s potential to benefit a variety of vertical markets such as automotive, farming, mining, construction, aerospace, transportation, manufacturing, and more. Welcome to HxGN radio. My name is Brian, and today we have Martin Rudigier, Head of Finance with Apex.AI with us to discuss how Hexagon technology supports innovation at Apex.AI. Martin, thanks for joining us.

MR: Thanks for the invitation.

BK: Of course. Appreciate it. And so first of all, let’s hear a little bit more about your company. Tell us what you do, all of that good stuff.

MR: At Apex.AI, we built an operating system for autonomous vehicles that never fails. A little bit about my background, I joined the company roughly one year ago as the first business hire, and we’ve since seen the company triple in size. So that has been very thrilling and exciting to see. Before Apex, I worked at Daimler, McKinsey, and Autotech Ventures, and I’ve seen the automotive industry change over the past 10 years. Then met Jan Becker, the co-founder and CEO of Apex.AI at a lecture while I was getting my MBA at Stanford.

BK: Nice.

MR: Shortly after, I joined the company on what I think is a great mission. If we want to talk a little bit about what that mission is, we were founded in 2017 and we developed safe and certified software for autonomous mobility as you mentioned. And to be a little more specific, our first product, which is set to be released by the end of this year, is called Apex.OS, and it is a safe and certified software framework and software development kit for autonomous mobility systems.

And because that sounds probably a little complicated, if you think about your iPhone, in case you have an iPhone, basically what we built for autonomous cars is the iOS layer.

BK: Got you.

MR: The operating system that runs on the hardware where third-party developers can then program applications that sit on top of the operating system, and in the case of autonomous vehicles, that would be everything, all the algorithms that actually make the car drive itself, that would sit on top of our product. We just make sure that the operating system layer doesn’t fail. Basically we make sure that your iPhone doesn’t freeze, which it does.

BK: Well, yeah, that’s very important.

MR: Yes.

BK: So, okay. First of all, you’re doing consumer and you’re also doing commercial, is that correct? Or mostly commercial?

MR: We are the commercial, we are a B2B company. Our product will most likely be in autonomous cars, but you as a consumer riding in it or buying one, you will not know that it’s Apex.AI inside, necessarily.

BK: Sure. Yeah. Okay. So that’s an AI. I’ll word it in a different way in saying that you’re doing it for vehicles that consumers will buy versus just commercial vehicles.

MR: Yes.

BK: Okay. And you brought up something interesting because, I have a friend who has a Tesla, the Model 3, he’s driving it and he says, because you mentioned it, the screen locks up or the phone locks up, his screen went out. Now the car is still driving, but he’s in the middle of a freeway driving 70 miles an hour, having to reboot his computer, which is a scary moment. Then the question is, well, what if the car freezes and locks up? So, what you’re doing is you’re trying to say, we’re trying to guarantee that this never happens.

MR: Yes.

BK: Excellent. So are you going to be working with, I mean I know you probably can’t talk about specific companies, but for example like a Tesla-type thing where someone can say, “I buy a Tesla” or “I buy some kind of an electric car” or whatever, you’re going to work with something like that, and you’re also going to work with commercial entities as well, essentially.

MR: Yes. The product is so low, is on such a low level in the autonomous driving stack that the applications can range from passenger vehicles to rideshare vehicles, autonomous shuttles to autonomous trucks, potentially also aerospace, potentially drones, potentially mining, agriculture are big applications. Basically anything that moves through the world autonomously and therefore presents a hazard or a danger to other humans. That is all safety-critical applications. And we address all of those use cases with our product.

BK: I like that. So because you’re a B2B, you might be the company that people have not “heard of,” but that everybody uses essentially and we just don’t know it. I like that. Well, cool. Well, thank you for explaining that. Now you mentioned: never fails.

MR: Yes.

BK: I’m curious about that. Explain that a little bit more.

MR: Yeah. So if you look at the autonomous driving space today, there’s a lot of companies out there that receive a lot of funding. And really most companies are focused on solving the problem, which is currently still a research and development problem of basically developing a system that works, but not necessarily a system that works and is compliant to the functional safety standard, which is called ISO 26262 in automotive. And basically that is where we come in, exactly at this inflection point. So once the R&D challenge has been solved, companies will need an operating system that can take them into production.

BK: Okay.

MR: And that is what we provide with Apex.OS.

BK: Excellent. But you mentioned, too, that your product never fails as well.

MR: Yes.

BK: In the very beginning. So I mean, really never fails? I mean, is it a hundred percent kind of successful, or is that what you’re striving for?

MR: It needs to be 100%, and we will have it safe and certified by one of the most well-known certification agencies, an agent in Germany called TUV NORD. And yes, to your point earlier, it needs to be 100% and not 99.9999 because if you are sitting in that car, 99.999 still means that at like a hundred thousand miles, there might be one system failure. And that cannot happen.

BK: Yeah, of course. So you’ve got fail-safes and backup after backup kind of thing.

MR: Yes.

BK: That’s incredible. Really. I’m all for that. I like that. Now this is obviously very much technology here. How is technology helping you in this scenario, giving you the solution or how is it even hindering it? Making things harder?

MR: Yeah. As I mentioned, there are hundreds of companies currently testing autonomous vehicles and trying to solve the complex technical challenge of making the car move by itself in any sort of environment.

And if you look at those companies today, you might ask what kind of operating system are they using today? And 80% of them, 95% of academia, use what’s called the robot operating system or ROS. ROS is open source. It has a very strong community behind it, and it is good for rapid prototyping and R&D. But for some of the reasons that we talked about already, it is not a 100% fail-safe, if we want to use that term. So it’s not hard real time. It’s not fully deterministic, which means that if you execute a certain application at one million times, every single time, the exact same thing happens. Think back to your iPhone. When you open your LinkedIn app, sometimes it opens right away. Sometimes the memory is allocated to all the other apps that you might have been running on your iPhone.

And so the system first needs to grab some of that memory from other applications and make it available to LinkedIn, which is why then the LinkedIn app takes longer to load. Exactly the same process would happen on an autonomous vehicle when. For example, you need to initiate a brake. You cannot wait for some memory to be reallocated. That is what we solve with Apex.OS. And so going back to the fact that 80% of companies and 95% of academia currently use ROS, we make it extremely easy to go from ROS to Apex.OS, basically to move from the research and development world to actually deploy and sell to customers. And that’s this transition. This is where we come in.

BK: So you’re going to the customers now, you’re moving them from ROS to your product, and you’re teaching them how to do it.

MR: Yes.

BK: How is that working? I’m assuming it’s a simple process just because it sounds like you’re doing a pretty good job with your product.

MR: In terms of where we stand as a company, we are set to release the first feature-complete version of Apex.OS at the end of 2019.

BK: Okay.

MR: And we currently have pilot customers that are already developing on top of Apex.OS, but we will not see it in deployment in actual cars that drive on the road for quite a while, because obviously you need to first start working with us, then your developers. If you’re a customer, need to start developing applications on top of it. And the lead times in automotive are relatively long, can take up to like five, six, seven, eight years from starting a project to seeing it on the road.

And so that will take a while, but we are already working with customers.

BK: Okay. And it’s going very well.

MR: Yup.

BK: Excellent. And it’s very interesting to see the development of this. There’s obviously the city regulations as well, state regulations that come into play. There are so many external factors. Like you said, the lead times, especially with motor vehicles is just very long. But it’s good to see that this is in progress and actually happening. And you always hear the concerns, well, is it really safer autonomously. It seems like it’s going to be. That’s great.

Have you had any issues keeping up with the new technology? You’re implementing a lot of new things or even integrating technology. Like you said, you’re going to other companies and you’re dealing with what they’ve done too. So has it been hard to keep up with it, and how have you overcome that?

MR: I mean, in general, this space is moving very rapidly, but I think some of the hype that we’ve seen in recent years is slowly coming to an end. Or people are waking up to the reality that it’s still, you know, autonomous vehicles that drive anywhere autonomously are still many, many years out. There are going to be certain use cases, for example, shuttle services or rideshare services within very limited geographic bounds where you can even today probably solve it, solve that challenge to 99.99%, but overall the big challenge in this industry is that no one really knows when it’s going to be ready. And everyone is chasing basically to make that happen. But it’s going to take many more years until there’s going to be ____ (11:35 – “wided”?) option.

BK: Sure. But it’s progressing. Right?

MR: It is progressing.

BK: Okay.

MR: And what’s good is there is a lot of money flowing into this space, just because global transportation, right, is such a big market. There’s not many markets that are that big. The opportunity is so large that investors are pouring hundreds of millions and billions of dollars into start-ups sometimes just based on an idea or the profile of the founders, which we see in our neighborhood in Silicon Valley almost every day.

BK: Yeah, of course.

MR: So that is a super exciting time to be in that space right now.

BK: That’s great. Yeah, that’s great. Well, how has Hexagon and their technology supported your innovation efforts?

MR: Within Hexagon, we’ve so far predominantly worked with AutonomouStuff, which was acquired in 2018 by Hexagon, and specifically we bought our own test vehicle from AutonomouStuff, including all the sensors and components. And we now operate that vehicle, it’s a Lexus RX 450 hybrid, basically as one way for us to test our own product. Again, we don’t build an autonomous vehicle, but we need to make sure that we are able to prove our claims in practice. But yeah, related to AutonomouStuff, we started out in 2017 and really worked with Bobby, the founder and CEO of AutonomouStuff from the very beginning. He’s in a very interesting position in the space just because he interacts with basically all players. I’m sure in Silicon Valley, almost every other start-up or maybe even more gets their cars from AutonomouStuff. And so he has a lot of insight into how this space is evolving and we’ve always found it super helpful to talk to him and see how he thinks about where the industry is headed.

That said, we are looking into deepening that partnership or this relationship, which was more like a customer relationship into more of a partnership, and integrate our software offerings with the hardware offerings of AutonomouStuff. Ultimately we would like to offer our safe, secure, reliable software products to AutonomouStuff customers who want to move from R&D projects to productization.

BK: Awesome. Well it’s great to hear that. All right. So I want to know, because what you’re doing is extremely innovative, very exciting stuff. What does innovation mean to you and where are you going with that as we move forward?

MR: I think about innovation as basically looking around you today, every single thing that you see was at some point some sort of innovation. Even fire was at some point an innovation. And so when you think about it that way, then not every innovation has really made our lives better.

And AV technology, autonomous vehicles, have so many great benefits, fewer road deaths, no more drunk driving, et cetera. But the push towards realizing that has also come at some costs, including the cost of human lives, for example, in the Uber accident. And that is why we at Apex are focused so much on safety and not just proving that we can solve the challenge but proving that we can solve the challenge safely. And in a way that doesn’t come at anyone’s cost. And so that is why our software is, we say that our software is not designed for peak performance, because it isn’t. It is designed to never fail. And that is exactly our differentiator too. Let’s say an iOS, which is consumer grade, not automotive grade, because it doesn’t need to be, no one cares if the LinkedIn app takes a little longer to load. But you do care if your car takes one or two more seconds to open a safety-critical application that will determine whether you live or die basically.

BK: Sure. Which is, I mean you’re gambling with someone’s life versus time.

MR: Yes.

BK: You know, in that situation. And I like what you said too, that some innovation has been incredibly beneficial for us and some hasn’t. And you also see the situation where a company will rush to get something out and they’ll put it out while it’s not entirely complete. And so you have bugs, and like you said, things fail, things take too long, and then you’ve got 40 patches. So getting it done right. Making sure that you’re not doing it, like you said, just for the money, just for the business standpoint, but for safety and actually being a benefit so that your innovation truly is beneficial.

MR: Yeah.

BK: I like that a lot. Well, thank you very much Martin. Appreciate you sharing all of this and appreciate your time today.

MR: Oh, thank you.

BK: It’s been great. For more information on today’s topic, please visit apex.ai. Go on over there and check it out. And of course, to learn even more and listen to additional episodes, head over to hxgnspotlight.com and thank you very much for joining us here on HxGN Radio.