F1 or Grand Prix (GP) Racing has over 400 million views over all Streaming & Television platforms. Nowadays, everyone has updates about all F1 Series around the globe which has clearly made a hype about this sport.
This was simply possible using Cloud Computing & Machine Learning Models. Director of F1 — Ross Brawn decided to partner up with Amazon Web Services (AWS) to maximize the scope of understanding of the match to users. This surely requires High Performance Computing (HPC) technology since these ML models work live when the race is in progress. The ML models simply collects data points from the race cars and predicts the future Pit Stop, Striking Distance of the car in the front, Driver & car performance, undercut threat etc.
Workflow & Architecture of AWS Tech designed for F1.
Stage 1: Data leverage
During each F1 race, these guys collect & analyze data from over 300 sensors across cars & the race track. These sensors approximately generate over a million data points every second that are leveraged to make decisions.
Stage 2: Leveraging the Inferences
All the predictions made from the transmitted data points AWS using HPC to improve performance of the car & its driver. Here AWS runs aerodynamic simulations to generate faster cars with lesser downforce loss impact. If I simply break it down in Driver and Fan perspective, it simply increases the chances of overtaking the racer in front for the driver as well as gives an exciting wheel-to-wheel action for the fans to watch and stay connected to the sport and later on which ends in social media popularity like being trending on twitter.
Stage 3: Improving audience experiences
F1 uses complex ML algorithms developed on AWS Sagemaker to analyze which camera view of which racer is exciting to watch at the very second for the audience to keep up the hype and make a decision accordingly. F1 simply uses old race recordings as the feature data to be given to predict better targets in order to engage the audience. The previous race recordings are analyzed and on the basis of those insights real-time decisions are made.
Architecture of AWS Cloud for F1
If you’re willing to learn more about how those algorithms are developed here’s a book by AWS themselves about it: https://pages.awscloud.com/GLOBAL-LN-GC-400-formula-1-ml-ebook-2020-learn
