Balance Control Ability Evaluation & Center of Pressure (COP) Classification with Machine Learning Methods

Document Type : Research Article

Authors

1 Advanced Instrumentation Laboratory, School of Mechanical Engineering, College of Engineering, University of Tehran

2 School of Mechanical Engineering, College of Engineering, University of Tehran

Abstract

Athletes' balance control ability is essential in different sports. Effective analysis of an athlete's balance control ability is a valuable solution for coaches and sports teams in identifying the skills of their subjects. In recent years, with the rapid growth of technology in sports, the need for intelligent methods has increased. This study compares various artificial intelligence approaches to assess balance control ability by analyzing time-series data of the center of pressure (COP). A recording pad collects COP data for four different types of subjects, from professional skiers to non-athlete individuals. Several experimental feature extraction techniques were implemented on the data, and features were utilized as inputs for artificial intelligence methods. This paper utilizes a multi-layer perceptron (MLP) to classify subjects' skill levels. Through comparison with other methods, MLP achieves an accuracy of more than 92% in classifying subjects' proficiency, demonstrating the best performance. Other methods, comprising k-nearest neighbor (kNN) and support vector machine (SVM), achieved 72% and 69% accuracy, respectively. Analysis of COP data can be useful in recognizing talented individuals who are promising for real-world applications.

Keywords

Main Subjects

AUT Journal of Mechanical Engineering
Volume 10, Issue 2
April 2026
Pages 231-242

Files

Share

How to cite

Statistics