Algorithms behind the Wheel: AI Impact on the Future of Motorsport
- Letizia Bizzotto
- Mar 25, 2024
- 4 min read
Episode 1: Data Analysis and Fan Engagement
In an era where the relevance of artificial intelligence in every field is unavoidable because it also influences public debate and information dissemination extending beyond technological boundaries, motorsport cannot consider itself exempt from this all-encompassing transformation.
In this context, the integration of technologies such as data analysis, simulations, aerodynamic modeling, fan engagement, and autonomous driving series are just a few of the areas where machine learning and deep learning are profoundly innovating the sector.
In the first episode of "Algorithms at the Wheel: The Impact of AI on the Future of Motorsport", the first column of Race Diary dedicated to Artificial Intelligence, we will explore two traditionally distant dimensions of racing: data analysis and public engagement.
Data Analysis and Strategy Simulation
This would be a mutilated sport without the vast amount of data and simulations it voraciously uses.
Quantitative analysis is a process that not only optimizes race strategies and car design by exploiting machine and deep learning algorithms to process data from hundreds of sensors on the cars but also feeds digital models capable of anticipating a series of possible scenarios in response to variables such as weather, competitors' behavior, pit stop strategies, track conditions, collisions, mechanical failures, and historical data collected in previous editions of the same GP.
Partnership with technological partners like AWS (Amazon Web Services), Dell, Oracle, and Ansys plays a key role in enhancing teams' analytical capabilities, improving not just technical performances but also strategic decision-making during races: Mercedes-AMG Petronas and TIBCO, Scuderia Ferrari and AWS, McLaren, and Red Bull with Dell, IBM, and Oracle, respectively. These are collaborations that provide teams with the tools to analyze a wide range of variables in real-time and select the most advantageous strategy based on race conditions.
In this context, where the impact of each variable can be predicted with greater accuracy than ever before, teams dedicate a large portion of resources to modeling millions of potential race parameters, with the goal of identifying those factors that most lean towards positive outcomes. These simulation cycles allow for identifying weaknesses and potential failure factors before hitting the track, a decidedly more economical and strategic solution considering the strict development and design budgets.
Infographics of circuit and car data flow © Amazon SageMaker
James Vowels, team principal of Williams, highlighted the unique role of AI in providing access to the valuable insights hidden within the large volumes of data generated during modern F1 races.
With a car generating up to 70,000 data channels in real-time, the challenge no longer lies in monitoring a few dozen channels, but in managing and interpreting this torrent of information:"They come across the screen in squiggly lines, and you look at them for patterns. The difference is, it's not 32 channels any more. There is no way, even with all the resources that we have, that you can consume them".
Vowels points out that the use of data science and machine learning is essential to compare data with simulations and identify anomalies or confirm trends, considering that cars undergo changes from one race to another, on different tracks and tires, therefore, in reality, the degree of commonality is quite small.
Fan Engagement and Cloud Insights
Since 2018, Formula 1 has initiated a collaboration with Amazon to innovate not only race strategies and data monitoring but also to revolutionize the spectator experience, aiming to make motorsport more accessible and engaging.
Ross Brawn, managing director of Formula 1, has set the goal of "bringing spectators closer to the dynamics of the pit stops, offering them access to data previously exclusive to race teams."
Each Formula 1 car is equipped with 100 to 300 sensors, generating and transmitting about 3-4 GB of telemetry data every 30-40 minutes, producing up to 1 million data points per second.
The innovation in fan engagement translates into the strategic use of these vast volumes of data, potentially abstract for the non-specialist public are made accessible and interpretable through targeted processing and presentation to enrich the visual experience, providing insights that stimulate greater involvement and interaction.
Formula 1, a notoriously complex sport, avails itself of the collaboration with AWS to analyze real-time positioning data and lap times, generating in-depth insights that are integrated into television coverage and live commentary, thanks to the F1 Insights platform. Utilizing the historical race data archive, stored on Amazon S3 and analyzed using advanced analytical models, Formula 1 shares with its fans detailed information that reveals the subtleties of decisions made in fractions of a second and highlights performances through explanatory statistics.
Finally, according to a report published by Nielsen Sports, Formula 1 has seen a 35% increase in its global fan base from 2019 to 2024, thus positioning itself as a precursor in the adoption of machine learning algorithms applied to sport, influencing other international leagues such as the NFL and the Association of Tennis Professionals (ATP), which have followed suit by adopting similar technologies to improve fan engagement.
Data Analysis and Fan Engagement have gone from being two sides of the same coin with little in common, to being more interconnected than ever, demonstrating how the advent of Artificial Intelligence is rewriting the rules of this sport, both on and off the track.
Commentaires