Professional and amature cyclists alike use power output (watts) as a primary measure of performance. The key power metric that is used to assess fitness, calibrate training levels, and pace races is Functional Threshold Power (FTP), which is a measure of a cyclists sustainable power over a period of time. More specifically, from a physiological perspective, a cyclists FTP sits at the point where the amount of lactate produced by their muscles is fractionally below their bodies ability to remove it. This is also known as Lactate Threshold (LT). Just below the LT, the body is able to balance energy supply versus demand. Whereas just above it, excess lactate will build up and performance will deteriorate. Accurately measuring FTP is important because it is used to optimise training levels and specific sessions can be targeted at a cyclists FTP with the goal of increasing it.
This project evaluates the application of machine learning for predicting an athlete’s FTP using data collected from previous training sessions. For more information please read the report.