Description

Abstract: Concussions pose a significant risk in sports, including kendo. This study aims to
use artificial intelligence (AI) to predict the concussion prevention potential of helmet inserts
with different properties. The primary research question was whether AI can accurately
predict the force reduction provided by various kendo helmet inserts, when compared to
actual physical testing. The insert pads were classified based on their difference between the
original thickness and the thickness when compressed under load. All the inserts tested were
cotton-based, commercial, off-the-shelf products. A random forest machine learning
algorithm was used to create the AI model. The study found that the AI-predicted average
force received by a kendo helmet insert was statistically indistinguishable from the actual
measured force (p-value = 0.38), with a 95% confidence interval that closely matches the
observed data. These findings suggest that AI can be a reliable tool for assessing the
protective potential of kendo helmet inserts, offering a cost-effective alternative to physical
testing. This study demonstrates the feasibility of using AI to predict the protective
performance of a kendo helmet insert, which could pave the way for future innovation in
sports safety equipment.

Keywords: Kendo, Random Tree Forest Model, Artificial Intelligence, Concussions, Sports
Safety