E/E Architecture Pioneer Ricky Hudi, a Decade Ahead, Talks Autonomous Driving

 

AEM interviewed Ricky Hudi, Chairman of The Autonomous. As Audi’s former head of electronics development who helped pioneer the end-to-end path of autonomous driving, he recounts the origins of innovation and many of Audi’s “firsts,” including zFAS and MMI. On The Autonomous, he shares the initiative’s philosophy - standardization and safety (especially AI) with fail-operational systems - and its 2.0 expansion into robotics, agri-tech, and health-tech. He also reflects on the significance of cooperation with the semiconductor ecosystem including NXP and Infineon, and with Samsung and LG, as the driving force for the SDV era. Finally, he offers a warm message to Korea’s auto industry: “When we join forces at the pre-competitive stage, 1+1 becomes far more than 2.”

By | Sang Min Han han@autoelectronics.co.kr
한글로보기





 
Before we dive into the evolution of end-to-end autonomous systems and The Autonomous initiative, let’s start with the early days. You’re a ‘star’ in the electrification of cars. How did your experiences at BMW and Audi shape how you lead The Autonomous today?
Hudi        I majored in electrical engineering, focusing on computer science and chip design. It was the early 1990s. After graduating, I decided to join BMW in Munich because, at the time, BMW was the undisputed leader in automotive electronics.
As a young engineer, I was passionate about in-house development. BMW was directly developing advanced electronic control units such as engine ECUs and gearbox control units. We were working at the cutting edge, based on 32-bit microcontrollers and a real-time operating system (RTOS). What I learned then profoundly shaped my entire career. I gained the capability to treat systems not as black boxes, but as white boxes - to fully understand and design what’s inside.
I later moved to BMW Technik, BMW’s internal research division, where we had great freedom to explore new automotive technologies. Around 1994–1995, we researched integrated HMI concepts, optical bus transmission, steer-by-wire, brake-by-wire, high-voltage systems, and computer systems. It was a truly enjoyable and challenging time.


How did you end up at Audi?
Hudi        I grew up in Ingolstadt, so I had always thought, “What if I work at Audi someday?” Frankly, in the early 1990s - at least until 1996 - Audi’s electronics lagged far behind BMW’s. Then change came in 1996. My former BMW boss, Willi Baschleiter, was recruited from Bosch to become Audi’s new E/E (electrical/electronic) leader. Through him, I first connected with Audi, and it felt like a very inspiring combination from the start. In 1997, I left BMW and moved to Audi.



At CES 2014, Audi led automotive trends by presenting the zFAS controller, enabling cars to drive and park themselves, personalized digital entertainment, and new uses for laser light.


 
Audi did extraordinary things in those years, didn’t it?
Hudi        At Audi, we built a new in-house advanced engineering team for electronics. Our mission was thrilling: to develop the electronic systems for the all-new Audi A8 (code D3) targeted for launch in 2002. Our goal was to close the gap with BMW and Mercedes-Benz in electronics within a single development cycle.
The project’s biggest success came in the fall of 2002, when Auto Motor und Sport compared the new HMI systems: our Audi MMI, BMW’s iDrive, and Mercedes-Benz’s COMMAND. The result was remarkable - Audi MMI ranked first and made headlines: “The Audi system delivers what the BMW system promises.” It was an unforgettable achievement for my team and me.
As for how my experiences at BMW and Audi influenced The Autonomous later: I learned how crucial networking is in this industry - building trusted, long-lasting relationships is key. I still have many friends at BMW. When I visited Bosch not long ago and met former colleagues, it felt like coming home. Those years were an important phase in my career and taught me the power of human relationships built on trust across the industry. That learning is deeply embedded in what I do at The Autonomous today.



Back then I wrote a lot about Audi - MMI, the digital cockpit, zFAS. Audi was at the forefront of electrification and a star at CES. How did you pull it off? 
Hudi        We need to start a bit earlier - with why we pushed all this development and what the inner motivation was. The essence is that I love innovation. My personal slogan was “innovation driven by passion.”
At Audi we delivered numerous firsts. I kept telling my team, “We must be first in multiple areas.” MMI was one, the modular infotainment platform MIB another. People may not remember this well now, but Audi was the first to bring NVIDIA into the car, via MIB in 2012. In 2014, we were the first to adopt Qualcomm Snapdragon in the car. We were also the first to integrate Samsung Exynos.
We also introduced Matrix Beam headlights to the world, allowing drivers to use high beams without dazzling others. And of course, the central driver assistance system, known in German as zFAS, was born with us. Looking back, we were a decade ahead of the industry in this domain. The system was already central-computer-based and had a multi-SoC architecture. In essence, it was what we now call a Level-3 safety architecture and a fail-operational system. Developing zFAS was a tough challenge, but it paved the way for the entire industry. Once we opened the road, many companies followed.
Ultimately, people matter most. I had the chance to hire top talent myself and build an A-team with A-players. I motivated them with a very clear vision - “Be first, be first again, and be first once more.”
We were also fortunate to showcase our work every year at CES starting in 2011. In those years, Audi was a true front-runner at CES, leading the industry by presenting our innovations and technologies to the world.


 
Along the way you also had strong ties with Korea. How did that begin?
Hudi        I can be very specific. I’d divide the story into two parts. First, during my time at Audi - especially from the 2000s onward - we had partnerships with major Korean companies such as Samsung and LG. After becoming the global head of electronics and software development at Audi, we intensified collaboration with LG and Samsung - displays and batteries with LG, and especially semiconductors with Samsung.
Audi launched the Audi Semiconductor Progressive Program, which enabled Audi to be the first to adopt Samsung Exynos in cars.
Then, in September 2016, I founded my own company. Right after, in October, Young Sohn (then CSO at Samsung Strategy & Innovation Center) reached out to consult me on Samsung’s acquisition of Harman and to connect with key stakeholders in the auto industry.
This ties back to TTTech and TTTech Auto. I discussed joining forces again with Georg Kopetz and Stefan Poledna (from the zFAS days), and in September 2016 we proposed separating the automotive business from the TTTech Group. Several things converged at that moment: my contact with Young Sohn, and Samsung’s plan to expand investments in automotive. This ultimately led to Samsung investing in TTTech Auto. Through Young, I built many good relationships and had several meaningful meetings with Samsung contacts in Korea.



When did you first meet TTTech Auto’s CEO, Stefan Poledna - was it during the zFAS days?
Hudi        From early 2009, I was responsible for Audi’s electronics development and also served as Vice Chairman of TTTech Group’s Supervisory Board. I worked very closely with Stefan Poledna and Georg Kopetz from then on. In 2011, I told Stefan and Georg, “Let’s think about how TTTech can become a true software partner for ADAS and AD for the Audi/VW Group.” That was the start of zFAS development. TTTech handled software integration for the project, and the version before MotionWise was called TT Integration. Through the zFAS development, we effectively entered the path that led to the MotionWise middleware. I’ve had a relationship of trust and collaboration with both of them from very early on, and that continues unchanged.


 
Could we say zFAS already had a Level-3 architecture back then? Many OEMs are tackling autonomous driving now, and as discussed at The Autonomous this time, safety architecture is a hot topic. What still needs to be done for AD ECUs? And how do you view the development speed of Tesla and China?
Hudi        First, we need to raise the level of standardization for safety-critical systems far beyond today’s status. Approaches are too diverse right now. Everyone is trying their best, but if we follow the principles set out in The Autonomous’ second report, we can save a lot of cost and headaches. So industry alignment on standardization is a key task.
Second, how to deal with AI and safety within fail-operational systems is still not fully solved. This too requires alignment and standardization.
As for Tesla’s development speed, I don’t want to judge that - it’s not my role. If we look only at what’s in the market, to my understanding Tesla does not yet have a Level-3 system on the market. If that’s the case, you can deduce the answer about development speed yourself.



The Autonomous has evolved - together with NXP, Infineon, TTTech, and TTTech Auto - from a “safety-focused, automotive-centric initiative” into a new organization redefining its mission and scope. 
From left: Susanne Einzinger, Head of The Autonomous; Stefan Poledna, CEO of TTTech Auto; Lars Reger, CTO of NXP; Georg Kopetz, CEO of TTTech; Peter Schafer, CSO of Infineon; and Ricky Hudi, Chairman of The Autonomous. In the background: members of The Autonomous ecosystem.







Many OEMs still face difficulties moving to Level 3. Over two days here, we’ve mainly discussed Level 4. Why Level 4? And what’s your view of open source in autonomous driving?  
Hudi        We intentionally put Level 4 at the forefront because, in reality, the boundary between Level 3 and Level 4 isn’t so clear. The real shift is between Level 2 and everything beyond, the point where responsibility for driving transfers from human to machine. Once you remove the human observation and decision maker from the loop, the entire system architecture changes. Call it Level 3, 3.5, or Level 4 - the game changer is removing the human, which means the architecture must be fully fail-operational.
So yes, many companies are moving toward L3/L4, but in a step-by-step approach: validate at Level 2++, then gradually transfer responsibility. Some systems are already on the road, but the approach is cautious and incremental.
On open source: from an architecture and safety perspective, whether it’s open source or not isn’t the decisive factor. What matters is whether the software is built to safety requirements. Even if it’s not open source, source access is necessary for safety reviews and inspections. So the difference is more about business models and collaboration modes. The The Autonomous initiative addresses how to design safe system architectures; the next step - implementation - can be in-house, via partnerships, or as open source. That’s the difference between what we do and other initiatives. The core is adherence to safety principles.


 
Tell us more about The Autonomous and your philosophy on autonomous driving. What does the industry need now? 
Hudi        I believe joining forces in this field is not only helpful but absolutely necessary, and I hope the industry realizes this soon. As The Autonomous moves into its next stage (2.0), I believe it can serve as a cross-industry catalyst for exactly that.
Let me cite two examples: telecom and semiconductors. In terms of standardization, these sectors are about 40 years ahead of automotive and have enabled many things. That doesn’t mean automotive should despair. On the contrary, this is an opportunity, because the meadow of opportunities here is still very high - there is vast, untapped ground.


What is this year’s KPI for The Autonomous? 
Hudi        The KPI is right here on the table, as Stefan showed: the second report published by our Working Group “Safety & Architecture.” They’re just getting started on a new phase - we shouldn’t miss that. With the second report, we expanded the candidate safety architectures, broadened the assessment methods, and expanded the applicable standards. We want to send a signal to the public and the industry: “This is a reference - read it and use it.”


At The Autonomous venue, other Europe-level initiatives such as HAL4SDV and Shift2SDV are being held alongside. How do you relate or differentiate, and what impact might such initiatives have? 
Hudi        In general, we are a fully open initiative and do not compete with others. That’s a clear and firm stance. We also don’t see overlap. What we put over the entire car - an end-to-end architecture - is comprehensive in scope and uniquely positioned. Our way is simple: anyone willing to contribute with good insights is highly welcome.
For example, the initiatives you mentioned - like Eclipse S-CORE, which deals with middleware - are good initiatives that I appreciate, because they align forces. But that’s one piece of the puzzle. We look at the entire scope and the entire picture. Another important point: we do not produce code. Our outcome is alignment on architectural principles at the pre-competitive stage. Competition begins after that, in how companies implement.


 
Founding members include NXP and Infineon. You started your career in chip design. What do semiconductor players mean to this initiative?
Hudi        Throughout my career I’ve always been connected to the semiconductor industry. Semiconductors are very important - almost every innovation in daily life relates to them. Whether it’s smartphones, medical devices, cars, security - whatever it is, semiconductor technology underpins it. Knowing how to interact with that industry and leverage it is crucial. That’s why companies like NXP and Infineon are not only excellent - they are absolutely essential to automotive and many other industries.


The Autonomous now includes robotics beyond automotive. What does the future look like? Will there be new working groups? 
Hudi        Absolutely. We’ve already expanded our scope to safety, security, and robustness through cross-industry collaboration, and we’ve formed a new organization for that.
In automotive, the first fail-operational systems have entered mass production, and the market has learned what’s possible and what isn’t - what’s the right path and the wrong one. This market learning is exactly the foundation that lifts us to the next level.
We’re ready to write the next chapter on the basic principles for building safe and trustworthy systems. We call this “The Autonomous 2.0,” and we’re pursuing it within a new organizational framework. We’re considering expanding into robotics - particularly humanoid robotics - as well as agri-tech and health-tech, because the foundations for scaling robotic systems are the same as for scaling autonomous automotive systems. As Lars Reger (NXP) mentioned, robots need human trust. What are the KPIs of trust? Safe, secure, and robust.
As for working groups, we’ll see how things develop, but if it makes sense, we’ll do it. We are very much looking forward to the momentum ahead.


 
Lastly, do you have a message for Korea’s automotive industry?
Hudi        I have deep respect for Korea’s auto industry. A former colleague of mine built a fantastic career there - Peter Schreyer. He was my colleague and Chief of Design. We once competed around the Audi D3 model.
What Hyundai and Kia have achieved is truly impressive - great cars, great technology. I think, like every other auto industry, Korea’s should recognize the value of collaboration. When we join forces, 1+1 becomes far more than 2. There’s plenty of room for competition, of course, but at the pre-competitive stage there are many problems we can solve together. That’s the basic idea of The Autonomous.



 

---

Bosch’s AI Declaration: A Future Beginning with Level 2++
Vision-only Self-driving Cars Will Never, Ever, Ever, Ever Happen!
안전을 설계하다: E2E 아키텍처와 자율주행의 미래
The Autonomous, 안전 아키텍처 로드맵 2판 공개

AEM(오토모티브일렉트로닉스매거진)

---