ItalyManufacturing2023 – ongoingFerrari
Pit-stop video and telemetry synchronisation for Scuderia Ferrari HP
Scuderia Ferrari HP cut pit-stop analysis to 60-90 seconds per stop with a green-light detection model, per AWS
Problem
Race engineers received about 22 pit-stop videos per driver each race weekend, roughly 600 a season. They synchronised the video with equipment telemetry by hand, and analysis took a few minutes per stop.
Approach
With AWS, the team trained a YOLOv8 object-detection model in Amazon SageMaker to find the green lights in front and rear pit videos. The lights mark the points used to align video with telemetry, and engineers review the result in a custom visualisation tool. Development began in late 2023, the first test was at the March 2024 Australian Grand Prix, and production use began at the April 2024 Japanese Grand Prix. For 2025 the team moved from two cameras to four.
Results
Takeaway
The team trained on 2023 race video and ran the system in production within about a season. AWS also said an anomaly-detection prototype for pit stops was in development.
Categories
- Industry
- Manufacturing
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer vision
- Adoption
- Vendor solution
- Evidence
- Vendor case study
Sources
Source summaries
The AWS blog explains how a YOLOv8 model on SageMaker detects pit-stop green lights to sync video and telemetry. It gives the development timeline and says synchronisation is up to 80% faster and analysis takes 60-90 seconds per stop instead of a few minutes.
Full details are in the original sources.
