Location technology

Combining telemetry data to build a positioning model for a racing line

Formula one, Sport

Determining the exact location of cars on the track at any given time is critical to supporting television broadcasts.

Download PDF

Goal

Formula One features 20 cars hurtling around race tracks several kilometres long at speeds of 375 kilometres an hour. Given those dimensions, locating competitors at any given moment can be a steep challenge, even with digital tools. F1 leadership sought to build the capability to be able to locate its cars on the race track in real time to within a few centimetres to support race management and broadcasts.

Insight and Action

F1 tracks have sensors positioned around the track that generate data continuously throughout the race, but the quality and reliability of these sensor streams can vary considerably. To address the data quality issues, QuantumBlack tested two separate probabilistic filtering techniques to combine the sensor streams. This approach allowed our team to evaluate source reliability, any periodic glitches, and the likely physical behaviour of the car. Our analysis provided F1 with the locations of each car that were within 5 centimetres of their actual position.

Results

  • 5Centimetreaccuracy target for image processing phase

Based on the success of this effort, F1 leadership subsequently expanded the project to incorporate additional image-derived positioning information.