"When AI is Used in the Right way it Will Make the Roads Very Safe" - Autobrains’ Joachim Langenwalter

Autobrains is an AI mobility company that's thinking differently to most autonomous driving companies. It is developing hardware-agnostic, self-learning artificial intelligence (AI) for advanced driver assistance and autonomous driving systems (ADAS and AD). The Israeli company is backed by major players such as Continental, BMW and VinFast.

It's also just announced that Joachim Langenwalter, former Stellantis Head of Engineering, is joining the company.

Langenwalter, Autobrains’ Senior Vice President of Autonomous Driving, has been talking to Auto Futures.

"I’ve been engaged in the autonomous vehicle industry since it began, and when I looked into what Autobrains is doing I could tell that the business case for the technology is amazing and is disrupting the industry. I’m excited to roll up my sleeves as part of a deep-tech automotive AI company," says Langenwalter.

At Stellantis, Joachim Langenwalter was responsible for the global automaker’s software and hardware engineering, including its autonomous driving, smart cockpit and connected services programs. Prior to Stellantis, he was Vice President of Software Strategy at Volkswagen, and Director of Automotive Software at NVIDIA.

After many years at OEMs, Langenwalter says he wanted to apply his knowledge to a startup again.

"What I have learned so far blows my mind. I have been in this industry for 30 years and what Autobrains accomplishes with its solutions is incredible. It solves the challenges the previous companies I have worked for couldn't, and Autobrains’ future is very promising," he says.

Autobrains AI doesn’t need to be trained by a human. Instead, it is developed to recognise patterns within the driving environment by itself .

"To make ADAS and AD solutions scalable and available in all markets, our approach is essential. We have seen many companies failing despite great investments and years of research because they were backing the wrong horse – supervised learning," explains Langenwalter.

According to the company, this unsupervised approach cuts the cost of AI for AD and ADAS by eliminating extensive data-labelling costs as well as significantly reducing in-vehicle chip and cloud processing and storage, therefore reducing overall hardware costs.

"It is almost technically possible to produce a fully autonomous vehicle with the current supervised learning technology, but it will not be economically viable or reliable. Which means that no automaker wants to produce it and a customer would not buy it due to its extremely high cost and safety gaps." 

"Another issue is that some automakers are stuck in a sunk-cost fallacy where they have spent so much money on AV technology that isn’t working – they’ve hit the supervised-learning wall – but they are unwilling to cut their losses and explore other possibilities," he adds.

M&A in the AV Sector

One of Autobrains' partners in the ADAS and AV marketplace is the Tier 1 supplier Continental, which is also a strategic investor in the company,

In a press release, Frank Petznick, Head of the Driver Assistance Systems Business Unit at Continental, says: "We are excited to be partnering with AutoBrains to bring to market its advanced and proven AI technology that we believe will disrupt the ADAS and AV marketplace. Historically, AV and ADAS technologies have been limited by their dependence on supervised learning that uses massive labelled training data sets and requires enormous compute power. AutoBrains' AI breaks through those barriers with a different approach that processes relevant signals from the car's environment in much the same way that human drivers do." 

Langenwalter believes collaboration is key for the AV industry's future development and commercial success.

"Collaboration is crucial, particularly in terms of different sensor inputs. Autobrains is currently developing systems that primarily draw input from cameras, and we have built great partnerships in that space. Knowing redundancies must be built in for safety’s sake, we are also looking into partnerships with radar companies and will eventually expand to include lidar companies as well. And of course, as a software company, we are building partnerships with other hardware manufacturers, Tier-1s, OEMs and cloud vendors for data storage and processing to build our programs into their products."

"Autobrains has built strong relationships with strategic partners and investors thus far, and our Chief Business Officer Hilla Tavor and I will both pursue additional opportunities. I believe we will also see collaborations take different shapes as M&A activity in the sector continues to play out," adds Langenwalter.

No Need to Fear AI in Automobiles

There's been a lot written about AI, particularly about the impact of chatbots, such as ChatGPT. There is also concern that AI may start replacing humans in the workplace. AI is already playing a leading role in the automotive industry, as ADAS functions become more prevalent.

"L2 and L2+ ADAS solutions are the stepping stones that will eventually lead to L5, true full autonomy. There are a lot of unknowns as we continue developing AI technology, and it will take a very long time until every single car in the world is fully autonomous, but AI has tremendous potential to make our roads significantly safer for everyone – and is doing so already today."

Finally, we asked Langenwalter whether AI should be feared, embraced or more tightly regulated.

"In the case of automobiles, AI solutions in general are trained for narrow scenarios pre-implementation in the vehicles. As soon as they are deployed in a vehicle that is being sold to an end consumer, the system does not learn nor grow anymore. Therefore, there is no reason to be feared. This is not generic AI that could be usable outside of the intended operation," he answers. "It is the other way around: when AI is used in the right way it will make the roads very safe." 

"On the legislation side, lawmakers will learn more about the capabilities and potential of AI and of course there have to be measures to ensure safety. The European Union in 2022 introduced its GSR (General Safety Regulation) governing ADAS, a first-of-its-kind piece of legislation that requires all new vehicles to have basic ADAS technology like auto emergency braking, driver attention monitoring, lane keep assist, and more installed as standard features by the summer of 2024."

"While the GSR only addresses L2 ADAS, it is a good first step toward laws that regulate fully autonomous vehicles. Our technology meets worldwide regulations and requirements such as the GSR and EU NCAP (5-Stars for NCAP 2023) and is backed by more than 250 patents, so we are confident that our approach will be embraced, first by the Tier 1s and automakers and then by drivers everywhere," concludes Langenwalter .