Visualisation of running form changes measured by wearable sensors for conditioning management, an application of the Functional Data Analysis

Abstract

Running is a widely-accepted activity among the general public, with runners aspiring to achieve optimal performance. However, established methods for the regular monitoring of running forms is lacking. To address this gap, we explore a versatile visualization method utilizing the widely-adopted Inertial Measurement Unit sensor. The running forms of 17-year-old male high school students were monitored during long-distance running training. Acceleration and angular velocity data were collected from a sensor attached to the lumbar region; data from the left foot contact to the next left foot contact were defined as the running cycle. Fatigue during running was assessed using the Borg Scale. The distribution of principal component scores obtained from functional principal component analysis of the running form data corresponded to changes in fatigue from one measurement session to another. However, no consistent trends or changes were observed across subjects. The running forms of participants who were measured twice exhibited a close distribution and similarity, yet unique features were also observed during each measurement. The findings suggest that changes and characteristics of runners' running forms can be readily visualized using a generic approach based on commonly-used sensors and functional data analysis.