In The Arena by TechArena - AI and Autonomous Advancement at Airbus

Episode Date: September 30, 2024

Join Arne Stoschek, VP of AI and Autonomy at Airbus Acubed, as he discusses the role of AI in aviation, the future of autonomous flight, and innovations shaping the industry at Airbus.

Welcome to the Tech Arena, featuring authentic discussions between tech's leading innovators and our host, Alison Klein. Now, let's step into the arena. Welcome in the arena. Welcome in the arena. My name is Alison Klein, and we're coming to you for a recap session from the AI Hardware Summit from last week. And we are with Arne Stoschek, VP of AI and Autonomy at Acubed, an innovation center of Airbus. Welcome to the program, Arne. How are you doing? Thank you, Alison. It's a pleasure to be here. I'm doing well. So why don't we just start with introductions? You've never been on the show before.

Tell me about A-Cubed and your focus on AI and autonomy. A-Cubed is an innovation center of Airbus. And at Airbus, we make things that fly. So that's everything in space, helicopters, and probably most known public is large commercial aircraft. If you fly as a commercial airline, that's about 50% probability that you fly in an Airbus airplane. That's where we focus the development of the AI and autonomy application. Now, obviously, you are producing some of the most advanced machines on the planet and the most complex machines on the planet, and you're integrating technology in every square inch of your aviation products. Can you provide a bit of introduction

about the history of your business and how it's become so technically focused. Let me start framing that with an understanding how big this industry is and how many people we transport and where it's going. So currently, this industry of commercial aviation transports about 4 billion passengers. So not billion, billion passengers a year. And basically, an Airbus airplane takes off and lands at the 1.3 seconds. And on average, you have basically the half the population from San Jose and San Francisco always in the air. So it's a very, very large industry. Now, where's it going? So within the lifetime of one product, which is about 15 years, we expect an increase in air travel and passenger volume from 4 billion to 8 billion. And 8 billion is the current world population. So what does it mean? You're in San Francisco downtown or New York downtown at rush hour on Friday, and now we double everything. So that gives an idea on just increase in complexity we face just handling this massive amount of growth. So that translates into a couple of topics, how do you manufacture the planes, how do you fly them, and then

also what are new business models and services that we can establish as a new revenue generator in this industry. And that's pretty much where we focus at this innovation center here in Silicon Valley with a specific focus on AI as one of the key drivers for decisions and analysis of data. Now, A-Cubed is something that I've obviously heard of Airbus, but I had never heard of A-Cubed before. Tell me about how the innovation engine works for Airbus and what the scope of focus is. Our focus is on AI application and autonomy. We started out in 2017 with the

development of what's commonly called an air taxi. That's a thing that flies. It's electric, but it could take off landing and fully autonomous as a demonstrator. And for me, coming from automotive and autonomous driving around, this was a super, super fascinating topic that brought me into Airbus. Then theCUBE morphed into building the groundworks for autonomy capabilities of the large commercial aircraft. And now we're basically leveraging this expertise that we built, such as for AI, to branch out in my little key strategic initiatives, which is what I already mentioned, the topic of autonomous flight, how to move that

from a would-have-could-have into the context of a future product. Then the topic, how can we apply AI and more intelligence in the context of manufacturing of aircraft? And then the third one we call Connected Aircraft Program is the engine data analytics recommendation systems, basically revenue generation with services in aircraft. That's pretty much a focus on, we looked at what is specific strengths of Silicon Valley, Airbus being a global company with this many different global engineering locations? And those areas are super strong in Silicon Valley. And that's the rationale for having this entity of Airbus here.

Now, the concept of autonomous has been around for a long time in aviation. We talk about it, putting things on autopilot all the time, and we take that for granted. But this area is evolving fast. So can you tell me a bit about how a3 is focusing on the advancement of autonomous in aviation? Let me start with autonomous and what you just said, automation. So if you look at the progression in the industry, pretty much with every generation of aircraft, there was a fundamental increase in automation. And in the main driver, what that is the increase in safety and the automation, I mean, the

many different definitions, it's pretty much you follow a set of prescribed procedures and the pilot oversees the proper functioning and steps in which needed. So with modern aircraft such as the Airbus A350, we have a very, very high degree of automation. There's a reason why the term autopilot comes from this industry. And the big, big step from automation to autonomy is that the aircraft as a machine is able to make decisions and that's a very different quality. So basically some of the decision-making responsibility is moving to the airplane and that is a new idea concept how we utilize a pilot. So the pilot is more of a higher level mission management instead of physical piloting of an aircraft.

When we look at what are the core technologies to enable that and also to help us understand what's the path to put it into a product context and how would we basically roll out this type of functionality in a future product. That makes a lot of sense. roll out this type of functionality in a future product. That makes a lot of sense. We're just coming off of the AI Hardware Summit where you gave a keynote. Why is infrastructure capability important to AQube's research in terms of advancement of AI and aviation? And what did you see at AI Hardware Summit that piqued your interest? The development and even more important, the testing and validation of AI-based functions. It boils down to a lot of things, but two particular things. Number one, the development and creation of this AI algorithms.

So we need to have some very, very powerful computing systems to train this type of algorithms. And second is how to utilize data. Specifically, the amount of data that we need for the development and testing of algorithms are basically orders of magnitude higher than what you would need for previous developments. And we're talk about multiple tens of petabytes that we need to have accessible and that requires a fundamentally different IT architecture or computing system

that we have to put in place, that we do have in place. That ties in the infrastructure topic and coming back to the specific conference, that is a very common topic for pretty much everyone who develops AI applications. And it's super interesting and important to see how this is done in different applications and related industries and what does it mean for us.

Part of being an innovation center for Airbus in Silicon Valley is also established partnership with local companies and basically increased the number of suppliers and partners that we have in Airbus, such as for this particular new topic. LESLIE KENDRICK Now, when you look at autonomous flight, what are the benefits for the aviation industry for its full manifestation, and what are the benefits for customers? There are several benefits. The one big concern that we have as a legal manufacturer and being part of this industry is what I mentioned,

scalability and safety. If you have such an increase in traffic density, you have to basically put more intelligence in that system, augment the human intelligence that you currently have, and give traffic managers or pilots the tools that they can handle this type of increase in density and with this comes this risk-prone information overload. So that is the benefit to the industry and the benefit to a customer. Everyone wants to fly.

There's an enormous increased demand for flying. And the industry basically needs to meet this demand and needs to be able to handle that demand. And what customers expect, rightfully expect, this is available, this is safe, and this is affordable. Those are exactly the drivers that we see for this topic. Arno, what are the use cases that you're targeting with AI that go beyond autonomy in aviation? Are there others that you see as high value for delivering the best products to your customers?

One is a topic that we are building up is the topic of services. You have a much better understanding on customer needs and what do customers want at a specific instance and how to translate it into specific service or offerings that we have. If you look at a typical internet company, they have a lot of those in Silicon Valley, the type of services they can produce and the type of customer benefit they can produce, larger insight on customers. What we're looking for is how we can translate that into our industry.

So if you're in a plane, pretty much it's a captive audience, what is it that customers would want besides just getting from A to B that would make the event of flying more interesting and more joyful? We look at data analysis topics, recommendation analysis, recommendation machines that are AI-based as a core technology foundation for this type of services. What key technology intercepts are you tracking to get you there across autonomy and services? It goes back to basically two topics. One is perception.

How can I perceive the environment and make sense out of it? And then the next topic is decision-making. So I have all type of external, internal influences. And what do we do next? The augmentation of intelligence, of human intelligence. And what is the specific challenge in our industry? It's we are a regulated industry and we're an industry that creates some

super complex systems like an airplane that has a very high degree of reliability and availability. And what it translates into, this is a super, super safe industry. And it's an industry that customers trust. If you compare the accident numbers between, say, automotives and commercial passenger traffic, you get an idea how different that industry is from others in terms of safety and reliability.

The key challenges or the excitement for an engineer as an intellectual challenge is how can we make this wealth of methods that are already developed in other industries, how to be applied in the context of our industry and how can we derive new topics or new applications and on top while actually increasing safety as a central paradigm of this industry. And so what we currently very much focus on is establishing both the means for massive data aggregation and processing, then the development processes for how we both develop and test algorithms and how we put it into a safety-weighted context, how to prove our safety.

And as I mentioned, that is a super challenging topic. No one has done it before. There is no blueprint to follow. And that is personally what makes this job super interesting. Arne, that puts so much poignant focus on the challenge ahead for you and your team. And I'm very glad that you guys are working on this. I think that this space is so interesting in terms of the deep connection with technology and the solutions that you deliver to the marketplace.

I can't wait to see what you guys do. Where can folks find out more about A-Cubed and connect with your team if they're like what they've heard today? Easiest way is to check out our webpage, a-cube.airbus.com. And we basically list a lot of topics and blogs that's really transparent for most of the topics

that we work on. And the idea is to disseminate the topics that we work on and generate this openness to work across different industries on creating solutions. That's so awesome. So, cube.airbus.com. Thank you so much for being on today.

It was so much fun to get to know you a bit and learn more about what's going on at A-Cubed. It was my pleasure. Thank you, Alison. Thanks for joining the Tech Arena. Subscribe and engage at our website, thetecharena.net. All content is copyright by The Tech Arena.