Pedal Power and the Leap to Formula 1
In just six years, AI-based startup Neural Concept has made a quantum leap from developing the world’s most aerodynamic bicycle to working with four out of 10 Formula 1 teams. Neural Concept’s core technology, Neural Concept Shape (NCS), is a machine-learning-based system that provides aerodynamic suggestions and recommendations, helping engineers avoid pitfalls and explore new directions in aerodynamic design.
The journey began when Neural Concept’s CEO, Pierre Baqué, was studying at a computer vision laboratory. He was approached by Guillaume DeFrance from the Université Savoie Mont Blanc’s IUT Annecy cycling team, who wanted to break the world record for bicycle speed. Baqué quickly developed a shape that resembled the current world record holder, resulting in what he claims is “the most aerodynamic bike in the world at the moment.” This bike is fully shrouded, with the cyclist enclosed in a composite cocoon, protected from the wind.
The success of NCS extends beyond racing. Neural Concept has secured contracts with global suppliers in the automotive and aerospace industries, including Bosch and Mahle. The automotive industry, in particular, values aerodynamics as it seeks to maximize the range of electric cars. NCS is also used to optimize battery-cooling plates, improving efficiency and extending range.
Formula 1: The Ultimate Laboratory
Formula 1 has become a popular sport worldwide, thanks in part to the Netflix series “Formula 1: Drive to Survive.” While the series focuses on inter-team politics, aerodynamics plays a crucial role in on-track success. Neural Concept’s software is helping teams like Williams Racing regain their competitive edge. Hari Roberts, Head of Aerodynamic Technology at Williams, explains that NCS is used to improve simulations and deliver better results in computational fluid dynamics (CFD). NCS significantly reduces the time required for CFD simulations, providing a competitive advantage within the strict testing restrictions of Formula 1.
However, the use of NCS comes at a cost. Baqué estimates that teams typically spend €100,000 to €1 million per year on NCS, a substantial commitment considering the $135 million annual cost cap for F1 teams. While Williams Racing cannot directly attribute lap time improvements to NCS, they acknowledge its impact on their car’s performance and correlation.
Beyond Aerodynamics
The potential of AI in Formula 1 extends beyond aerodynamics. There are discussions about artificial agents making race strategy decisions and optimizing car setups. The exponential growth of the AI/ML industry presents both opportunities and challenges for teams. Deciding which new tools to explore and adopt is a crucial task.
Neural Concept is also expanding its presence in the non-motorsport side of the automotive industry. They are working on developing more efficient electric motors, optimizing cabin heating and cooling, and even crash testing. While NCS can currently only simulate crashworthiness for individual components, Baqué sees potential in optimizing a car’s crashworthiness while reducing unnecessary weight.
As Neural Concept continues to innovate and explore new applications for AI in engineering, the advancement of AI supercomputing platforms could further push the boundaries of what is possible in the field.