AWS Car Learning Supports Pit Scuderia Ferrari HP Analysis

As one of the fastest sports in the world, almost everything is a race in Formula 1® (F1), and even the pit stops. F1 executives should stop the tires to change or make repairs for the damage suffered during a race. Eachdo the precious tenth of the second the car is in the pit is the lost time in the race, which can mean the difference between the podium or the loss of the championship points. The crews of the pits are trained to operate with optimal efficiency, though their performance measurement has been challenging so far. In this post, we share how Amazon Web Services (AWS) is helping scuderia Ferrari HP develop more accurate pit stop analysis techniques using machinery teaching (ML).

PIT STOP CHALLENGES Performance analysis

Historically, analyzing the PIT Prohibition performance has requested from dirty operations engineers to accurately review the hours of the cameras located at the front and back of the pit, then relate the video with the car’s telemeter data. For a typical racing weekend, engineers receive on average 22 videos for 11 potholes (per driver), reaching about 600 videos a season. Along with being time, reviewing the images is prone to inaccuracies. Since the implementation of the AWS solution, trail operations engineers can synchronize data up to 80% faster than manual methods.

Modernization through partnership with AWS

Partnership with AWS is helping Scuderia Ferrari HP modernize the challenging process of Pit Stop analysis, using cloud and ML.

“Previously, we had to manually analyze numerous video recordings and telememetry data separately, making it difficult to identify inefficiency and increasing the risk of losing critical details. With this new approach, we can now automate and centralize the analysis, gaining a clearer and more immediate meaning of any stops, helping us detecting errors and detecting us. our. “

– Marco Gaudino, Architect of Digital Transformation Application

The solution uses the discovery of the object located in the Amazon Sagemaker to synchronize the video capture with telemeter data from Pit Crew Equipment, and the architecture directed by the server optimizes the use of calculated infrastructure. Because Formula 1 teams must respect the strict budget and calculate the resource lids imposed by FIA, AWS demand services help Scuderia Ferrari HP avoid expensive infrastructure above.

Innovation driving together

AWS has been a partner of the HP Scuderia Ferrari team, as well as Scuderia Ferrari HP Cloud, Cloud Machine Learning Cloud, and Cloud Intelligence Cloud provider since 2021, partizing in Power Innovation inside and outside the trail. When it comes to performance competitions, AWS and Scuderia Ferrari HP regularly work together to identify areas for improvement and building new solutions. For example, these collaborations have helped reduce vehicle weight using ML by implementing a virtual land speed sensor, regulated the process of assembling the power unit and accelerated the prototype of new commercial vehicles.

After the start of development at the end of 2023, the Pit Stop solution was first tested in March 2024 at the Australian Grand Prix. It quickly moved to production at the 2024 Japanese Grand Prix, held April 7, and now provides the HP Scuderia Ferrari team with a competitive advantage.

Taking the solution one step further, Scuderia Ferrari HP is already working on a prototype to detect abnormalities during the pit stops automatically, such as the difficulties in raising the car when the carriages fail to remove, or issues during the installation and removal of tires from the Crew Pit. Also also setting a new, more performance configuration for season 2025, with four cameras shooting 120 frames per second instead of the previous two cameras shooting 25 frames per second.

Development of Solving Analysis of ML’s Ml Pit

The new solution of Pit stop with power ML automatically links video progress with telemeter-related data. It uses the discovery of the object to identify the green lights, then exactly synchronizes video and telemetry data, so engineers can review the synchronized video through a custom visualization tool. This automatic method is more efficient and accurate than the previous manual approach. The following image shows the discovery of the green light object during a pit stop.

“Systematically revising any stop, we can identify models, even discover the smallest inefficiency and refine our processes. Over time, this leads to greater durability and reliability, reducing the risk of errors that may endanger race results,” Gaudino says.

To develop the Pit Stop analysis solution, the model was trained using videos from the 2023 racing season and the Yolo V8 algorithm for identifying the object in the Sagemaker through the Pytorch frame. AWS Lambda and Sagemaker He are the essential ingredients of the Pit Stop analysis solving.

The workflow consists of the following steps:

1. When a driver performs a pit stop, the front and rear stop videos are automatically pushed to the Amazon simple storage service (Amazon S3).
2. From there, Amazon Eventbridge calls the whole process through various lambda functions, causing video processing through a multiple Queue Simple Raye (Amazon SQS) system and lambda functions that execute the personalized code to handle critical business logic.
3. These Lambda functions draw the right time from the videos, then join the front and back videos with the number of video frames containing green light to finally match the video joined by cars and racing telemets (for example, screw gun behavior).

The system also involves the use of the Amazon elastic containers’ service with numerous microsarvices, including what integrates with its ML model into the Sagemaker. Previously, to manually connect the data, the process lasted a few minutes per stop. Now, the whole process is completed in 60-90 seconds, producing close time knowledge.

The following figure shows the solution of the solution architecture.

cONcluSiON

The new Pit Stop analysis solution allows for a quick and systematic review of each stop to identify the models and refine its processes. After five races in the 2025 season, Scuderia Ferrari HP recorded the fastest stop in each race, with a better 2 -second flat season in Saudi Arabia for Charles Leclerc. The diligent work coupled with the ML strengthened solution more efficiently take the drivers again on the right track, focusing on achieving the best possible result possible.

To find out more about the construction, training and setting of ML models with fully managed infrastructure, see the start with Amazon Sagemaker. For more information about how Ferrari uses AWS services, refer to the following additional resources:

About

Alessio Louis It is an architect of solutions in AWS.