Algorithms behind the Wheel: AI Impact on the Future of Motorsport

Episode 2: CFD and Autonomous Series

As emerged in the first episode of our column, the concept of speed plays a crucial role in both motorsport and the field of artificial intelligence. Both domains are characterized by three fundamental aspects related to speed: the speed of transferring something X between two points a and b; the speed of transforming X into something else Y; and the speed of obtaining X, which is the time required to satisfy a specific need.

Having explored how these dimensions of speed have been applied in the context of strategic simulation and data analysis through the use of AI algorithms, we are now set to investigate the use of Computational Fluid Dynamics (CFD) and autonomous driving. These areas represent two more fields where the evolution of machine and deep learning is having a significant impact.

CFD simulation of an F1 Car © Grab Cad

Computational Fluid Dynamics

In Formula 1 CFD serves three roles: it is essential in the design process of new vehicles to evaluate the aerodynamic performance of updated components, for analyzing overall aerodynamic performance, and for identifying and solving issues when vehicles are not operating at optimal regimes.

Despite CFD requiring significant computing power and the expertise of qualified specialists, the benefits in terms of time savings and cost reduction far outweigh the initial burdens.

CFD simulation of a front wing © Grab Cad

For aerodynamic engineers, the main goals are to maximize aerodynamic downforce to optimize grip and maneuverability in turns, and to reduce aerodynamic drag to enhance straight-line speed, maintaining a balance that ensures consistent performance throughout the vehicle's operation.

In the dynamic context of Formula 1, aerodynamic analysis leverages a vast amount of data, from the track, wind tunnel, and simulator. To manage and interpret this abundance of information, the support of AI and HPC (High Performance Computing) based software is indispensable, capable not only of categorizing data but also of processing it to formulate effective race strategies.

This includes real-time analysis of variables such as tire pressure and brake overheating, which can significantly influence lap time.

The uniqueness of machine learning models and neural networks in this field requires that programming is entrusted to Data Engineers who not only have computer skills but also a deep understanding of the domain of expertise, namely aerodynamics. Only individuals with knowledge of both domains are capable of "teaching" machine learning systems the necessary aerodynamic principles, thus making machine and deep learning valid tools in service of performance.

Autonomous Series: The Frontier of Automatic Driving

The Indy Autonomous Challenge and the Abu Dhabi Autonomous Racing League are two of the most cutting-edge initiatives in the emerging field of autonomous automobile racing, challenging the current boundaries of artificial intelligence, electronics, and robotics applied to competitions.

During this year's second edition of the Indy Autonomous Challenge, universities and research centers are invited to design and develop Dallara AV-23 single-seaters driven by control software that, using sensors, supercomputers, and actuation hardware, allows the car to drive autonomously in the prestigious Indianapolis Motor Speedway arena. The software is programmed by students and researchers from the world's leading universities and research centers, including Italy's Politecnico di Milano and the University of Modena and Reggio Emilia.

Presentation of the car for 2024, the IAC AV-24 © Indy Autonomous Challenge

Concurrently, the Abu Dhabi Autonomous Racing League stands as another significant stage for testing the latest autonomous driving solutions, launching the initiative on April 28, 2024, at the Yas Marina circuit in Abu Dhabi.

Elite teams from prestigious institutions, including the University of California, Berkeley; Technical University of Munich; and Nanyang Technological University, Singapore, will work to automate the Dallara Super Formula, the SF23.

The event highlights international priority themes such as the future of transportation, inspiring the next generation of STEM talents, and ensuring a tangible impact beyond the track.

Faisal Al Bannai, Secretary General, Advanced Technology Research Council said:

"Abu Dhabi is a growing hub for STEM empowerment and for the vision of a decarbonised economy, which is why we are proud to launch the Abu Dhabi Autonomous Racing League, characterized by an open development model, supporting faster progress and testing with machine and reinforcement learning that will be critical to data collection and technology development for these vehicles".

Official Presentation of Abu Dhabi Autonomous Racing League © A2RL

In conclusion, the intersection created by machine and deep learning between the aforementioned spheres marks a pivotal technological turning point in the history of motorsport. The convergence of these developments not only redefines the concept of competition but also lays the groundwork for future developments in the automotive sector, from transportation to commercial, promising more efficient, safer, and potentially more sustainable vehicles.