Journal Article - International Journal of Performance Analysis in Sport
Roger Bartlett - 03 Apr 2017
https://www.tandfonline.com/doi/epdf/10.1080/24748668.2004.11868299?needAccess=true
This article looks at the development in the application of artificial intelligence to sports techniques over the years (roughly a decade) and how it has gone from not being used/ relatively unsuccessful to modern days uses for it in sport and how it might evolve and contribute to sports biomechanics in the future.
It mainly explores how artificial intelligence methods like expert systems, artificial neural networks (ANNs), and evolutionary computation have been applied to sports technique analysis and how these tools can evolve in the future.
It is outlines how AI can be used for diagnostic and predictive analysis of player techniques, helping coaches understand technical consistencies and weaknesses before competition. Bartlett notes that expert systems could “act as diagnostic tools for evaluating technique errors’” using rule-based reasoning. These system would use prior match data to automatically predict how player mechanics may perform under match conditions.
This would allow coaches and analysts to tailor the pre match training and tactical strategies based on individual or team-level biomechanics efficiency.
Bartlett highlights that AI systems such as ANNs can handle real-time pattern recognition and complex movements. The tools used process noisy, incomplete or fuzzy data to evaluate player technique consistency during play. The paper also explains how AI models “are good for classification, clustering and prediction tasks” and can “learn by experience”. This would support feedback systems and highlight technical degradation due to fatigue or tactical shifts.
At the minute this is not integrated but if it was, it would enable in match decisions support, alerting coaches to performance drops or suggesting substitutions and tactical recalibrations in real-time.