USING PARTICIPANT DESCRIPTIVE CHARACTERISTICS, HEART RATE AND INERTIAL SENSORS TO ESTIMATE RUNNING ECONOMY

Authors

Krista Vohlakari, Neuromuscular Research Centre, Faculty of Sport and Health Sciences, University of Jyväskylä, Jyväskylä, FinlandFollow
Sailee Sansgiri, Neuromuscular Research Centre, Faculty of Sport and Health Sciences, University of Jyväskylä, Jyväskylä, FinlandFollow
Timo Rantalainen, Neuromuscular Research Centre, Faculty of Sport and Health Sciences, University of Jyväskylä, Jyväskylä, FinlandFollow
Esa Hynynen, Finnish Institute of High Performance Sport KIHU, Jyväskylä, FinlandFollow
Neil Cronin, Neuromuscular Research Centre, Faculty of Sport and Health Sciences, University of Jyväskylä, Jyväskylä, Finland; School of Sport and Exercise, University of Gloucestershire, UKFollow

Category

Wearable Technology

Document Type

Paper

Abstract

In this study, we aimed to develop a machine learning algorithm to estimate running economy. Inertial sensor, heart rate and participant descriptive characteristics were used to estimate running economy by training and comparing five different machine learning models. Fourteen subjects performed a VO₂max running test on a treadmill with four-minute stages. Submaximal running speeds (lactate level < 4 mmol/l) were used to train the models. The best-performing model was a k-nearest neighbour regressor, which achieved average root mean square error of 0.097 ± 0.059 kcal/kg/km and average mean absolute percentage error of 8.7 ± 5.8 % compared to ground truth running economy data. Despite reasonably accurate running economy estimates, the model is currently not very generalisable, probably due to the small dataset used for training.

Recommended Citation

Vohlakari, Krista; Sansgiri, Sailee; Rantalainen, Timo; Hynynen, Esa; and Cronin, Neil (2024) "USING PARTICIPANT DESCRIPTIVE CHARACTERISTICS, HEART RATE AND INERTIAL SENSORS TO ESTIMATE RUNNING ECONOMY," ISBS Proceedings Archive: Vol. 42: Iss. 1, Article 191.
Available at: https://commons.nmu.edu/isbs/vol42/iss1/191