VDA Technical Congress 2022: Major Advances in Autonomous Driving

The Technical Congress in Berlin is always a bit like truffle hunting or gathering information in the Berlin metropolis. One repeatedly discovers hints at trends and new developments that coalesce over the course of the event as exciting sidekicks – and sometimes, they turn out to be merely politically correct but rather content-less platitudes.

Autonomous driving, however, is a different story. Richard Damm, President of the Federal Motor Transport Authority, introduced the most important facts. Germany has already enacted a national law on the matter, which could also serve as a blueprint for EU-wide legislation. This was further specified on July 28, 2021, for automated driving up to Level 4, creating three levels: 1. Starting with the operation of shuttle systems, usually in “closed” or better-defined areas, 2. The operation of retrofitted enhanced vehicles of all kinds before nationwide vehicle testing is expected to begin with Level 3 everywhere.

The laws have been passed, but full-scale operation is not yet possible. However, Damm appealed for support of this project, otherwise, Germany may lag behind, similar to the situation with e-mobility. The goal should rather be to set the global benchmark. After coming into force, Damm hopes for swift operational implementation, especially in defined operational areas, hub-to-hub services, and not just autonomous parking.

Dr. Joachim Bühler, CEO of the TÜV Association e.V., explained how everything must be safely accompanied and reviewed. He reiterated:

“There is a small but crucial difference between autonomous and automated driving!”

The former already exists, while the latter is the “big goal” towards accident-free driving. The expectations are accordingly high: too many people still die on the roads every day! And 90 percent of accidents are due to human errors. But the technology is also not error-free, as some sometimes fatal accidents with autonomous vehicles in the US – sometimes despite an overseeing driver at the wheel for emergency interventions – show. Hence Bühler warns:

“We must be careful not to replace human errors with technical ones.”

In this context, the latest surveys are also interesting: “The further we progress, the lower the trust in autonomous driving,” Bühler explains – and this is a global trend.

A completely new topic is also software updates: changing the car continuously during operation, for which Bühler chooses the performance increase as the most extreme update and jokes:

“Today, you get a performance boost through software, for which you used to have to tinker forever with your moped and then hope that the TÜV inspector would not notice the changes.”

Therefore, we must move away from spot checks to monitoring. But it's not just about software; new hardware like sensors, lidars, or cameras is also needed. And suddenly, a parking bump, a partially repainted bumper, or a crack in the windshield becomes a problem. To obtain precise data here, a medium-term digital vehicle registry and a legally secure path for insurers for how the software is deployed are needed. Bühler also emphasizes a crucial approach: Test first, then deploy. Because from the tests in the US, it is clear: Each individual accident does the technology a disservice! Where people tend to think software-wise and like to use customers as beta testers. But here it is about life and limb, which is why Bühler clearly demands: Test first, then implement.

From theory to practice: Where autonomous driving is already a reality today

How autonomous driving is further developed by suppliers was subsequently explained by Frank Petznick, Head of the Autonomous Mobility business unit and member of the Automotive Board, Continental AG. Although they come from hardware, they currently mostly hire software colleagues. Interesting are Conti’s observations in the practice of automated driving: First, the penetration is still not dramatically high, and the beeping systems are often still switched off. But they are getting better and more reliable with the development of levels three and four.

Currently, Conti is also scaling hardware and sensors for this purpose. According to Petznick, it is a significant investment and a “certain bet on the future,” especially since the return on investment comes significantly later. This requires new business models – which is why earning from software is also important. For this, cooperation is also needed. Petznick currently sees the implementable use cases (still) primarily in parking garages and on long-distance routes.

First Use Cases in Parking Garages

Dr. Eckard Steiger, Director of Industrial Cooperations Automated Driving at Robert Bosch GmbH, followed. He also emphasized the need for collaborations, along with a strong software background featuring powerful central computers and cloud servers. On this note, data quality and security are crucial topics. Accordingly, Bosch is advancing data-driven software development and cooperating in various practical applications, for instance, with Apcoa and Mercedes on Automated Valet Parking in Stuttgart, while testing with Ford in Detroit. These initiatives involve video-based systems with cloud services in the background, including for payment processing. Steiger also highlighted that they are starting with comparatively low speeds within a clearly defined and delineated space. He supports the VDA's lead initiative on setting regulatory basics for industry standards. Together with the so-called "AI Project Family," a consortium of multiple startups and industrial companies, they are investing 250 million euros into this area alongside the Ministry of Economic Affairs. This is crucial to remain competitive and to support fundamental research and standardization, such as the ISO PAS 8800 for road vehicles and artificial intelligence.

To tackle all these challenges, Steiger emphasized the importance of partnerships. Such collaborations can help share risks, reduce investment costs, and accelerate development. He explicitly welcomed additional co-developers.

What does this mean?

Automated and, in the future, autonomous driving is gaining momentum again, with Germany aiming to be a global leader. After a quieter period on the topic, recent legislative foundations and developments are propelling it forward once more. However, this is too complex for a single manufacturer to handle alone: Automated and especially autonomous driving requires strong collaborations.