How AI is penetrating the automotive industry

"AI will not only impact the vehicle but the entire company," emphasizes Jan Wehinger, Partner at IT consulting firm MHP. However, a major part of the transformation is not just technological innovations but, above all, the business mindset. The central question when implementing new technologies like artificial intelligence must always focus on their meaningful application in business.

Chinese customers are particularly interested in AI

MHP has examined how this can look in practice and under what conditions automakers, suppliers, or mobility players can make effective use of AI. The findings: Customers in China are significantly more open to AI solutions than those in the U.S. or Europe. Accordingly, their willingness to pay is also higher—although at a low level. While some customers are willing to make one-time payments when purchasing a vehicle, the vast majority prefer to use such systems for free. To increase the willingness to use or pay for AI assistants and to fully exploit existing potential, the industry must precisely understand which systems its customers want, says Wehinger. Particularly popular in this context are features that enhance the user experience, such as predictive maintenance or optimized route planning systems. When it comes to assessing benefits, the complexity of the solutions is of secondary importance, Wehinger explains. Instead, the goal should be to simplify everyday life for customers or save them time. Consequently, technologies like AI must always be approached from this perspective: The technology should not be used for its own sake but must provide a solution to pre-identified pain points or customer demands. The good news for automakers: "In the end, it will be about trust – especially with AI," says Wehinger. In this regard, OEMs are rated more positively by customers than many other companies or industries.

How ZF uses AI solutions

How deeply automakers and their suppliers have already integrated AI is demonstrated by Tobias Ehlgen, Head of AI for Systems and Control at ZF Friedrichshafen. The supplier uses artificial intelligence in areas such as quality assurance in production, autonomous driving, predictive maintenance, and natural language processing.

As a concrete example, the ZF expert presents a use case in trajectory control at the automotiveIT car.summit. Using the software cubiX.AD for autonomous driving functions, the supplier calculates the vehicle’s trajectory, based on which the relevant actuators are controlled. In this context, the supplier has developed the Safe BO algorithm in collaboration with RWTH Aachen, which provides advantages for executing specific parameters and optimizing the PI controllers in the vehicle. Through reinforcement learning, it is even possible to reach a point where the deployed neural networks make PI controllers in the vehicle obsolete and take over control themselves.

A second use case involves implementing the Eco Control 4 ACC predictive distance assistant, which uses map and sensor data as well as traffic sign recognition to create a horizon of approximately 500 meters. This enables more energy-efficient driving maneuvers to be planned. By utilizing AI, ZF has significantly optimized the respective vehicle systems and thus reduced the necessary resources. This ability to optimize even systems that are not inherently based on AI and to achieve faster embedded runtime is just one of the advantages of AI technology in the automotive industry. Additional benefits include faster development, better performance, and even safety improvements — provided that appropriate safety constraints are integrated into the systems.

This article was first published at automotiveit.eu