Related Experiment Video
Updated: Mar 31, 2026
06:52
An Inertial Measurement Unit Based Method to Estimate Hip and Knee Joint Kinematics in Team Sport Athletes on the Field
Published on: May 26, 2020
8.8K
Classification of team sport activities using a single wearable tracking device
Daniel W T Wundersitz1, Casey Josman2, Ritu Gupta2
1Centre for Exercise & Sports Science, School of Exercise & Nutrition Sciences, Deakin University, Melbourne, Victoria, Australia.
Journal of Biomechanics
|October 17, 2015
Summary
Wearable trackers can accurately classify team sport activities using accelerometer and gyroscope data. The Logistic Model Tree algorithm achieved the highest accuracy, demonstrating feasibility for sports analysis.
Area of Science:
- Sports Science
- Biomechanics
- Wearable Technology
Background:
- Wearable tracking devices with accelerometers and gyroscopes are common in sports analysis.
- Limited research exists on classifying common sports activities using these devices.
Purpose of the Study:
- To assess if a single wearable device can classify team sport activities.
- To evaluate the effectiveness of different machine learning algorithms for activity classification.
Main Methods:
- Seventy-six participants performed simulated team sport activities in a lab.
- Data from a MinimaxX S4 device (accelerometer, gyroscope) was collected at 100Hz.
- Features were extracted and screened; Random Forest, SVM, and Logistic Model Tree algorithms were used for classification.
Main Results:
- The Logistic Model Tree (LMT) algorithm achieved the highest classification accuracy (79-92%).
- Random Forest (RF) and Support Vector Machine (SVM) algorithms showed lower accuracies (32-43% and 27-40%, respectively).
- Feature selection reduced processing time but could impact accuracy; movement capture duration had minimal effect.
Conclusions:
- It is feasible to accurately classify team sport activities using data from a single wearable tracking device.
- The LMT algorithm demonstrates superior performance for this classification task.
- Balancing classification accuracy and processing time is crucial when implementing feature selection methods.
