Automated Force-Velocity Profiling of NFL Athletes via High-Frequency Tracking Data

Kevin Lyons completed his Master’s of Engineering thesis in Computer Science.

Abstract

The ability to measure key physical parameters of athletes is becoming increasingly critical for today’s sports organizations. Force-velocity profiling is a well-understood and studied technique for measuring the relationship between speed and output force in sport-specific contexts. Accurate force-velocity profiling systems can enable a wide variety of applications for sports organizations to improve player performance, cater better training programs, and potentially reduce injury rates in the long term. A current limitation of many of these systems is that they can require context-specific testing that impacts workflows for players, coaches, and trainers. Given the recent rise of wearable sensor technologies that track player movement in dynamic contexts, there is a clear opportunity to leverage new data streams to enhance this process. We present a novel system for automated force-velocity profiling using publicly available high-frequency tracking data of NFL players. We demonstrate that our derived force-velocity envelopes match observed position and player performance, and provide a proof of concept framework that would allow teams to leverage automated force-velocity profiling in their internal operations.

Read more here.