Skating Stride Algorithm

In hockey, groin injuries are prevalent and can often be attributed to adductor-abductor muscle strength imbalances. Quantifying leg load during skating is a first step towards identifying this imbalance. In order to quantify leg load, Catapult developed an algorithm to predict linear and crossover sections of skating using machine learning techniques based on data from an OptimEye S5, Catapult’s marquee wearable device. In the identified sections, we found individual strides from which load is calculated. We predict linear skating and crossover sections with an accuracy of 92%. Additionally, stride count is predicted with an accuracy of 92%. We speculate the error rate is from misidentified sections of linear skating and crossovers.

View this research