Document Type : Review Paper


1 PhD candidate, Department of Sport Biomechanics, Faculty of Sports Sciences, university of Mazandaran, Babolsar, Iran

2 Associate Professor, Department of Sport Biomechanics, Faculty of Sports Sciences, University of Mazandaran, Babolsar, Iran

3 Associate Professor, Faculty of Mathematics, University of Mazandaran,, Babolsar,, Iran


No systematic studies have been conducted to categorize the methods by which the ground reaction force could be estimated indirectly by using kinematic data and machine learning algorithms. The current review has collected and analyzed the studies conducted in the field of ground reaction force estimation using kinematic data during running from 2010 to 2023. These articles were found by searching the websites of Scopus, IEEE Xplore, Medline, ScienceDirect and PubMed using the keywords of Ground Reaction Force, Kinematics, Kinetics, Camera, Video Analysis, Accelerometer, IMU, in field, Running and a combination of them. 213 articles were found in different sources that after adapting to the inclusion criteria, finally 17 studies were reviewed. The review of articles showed that the indirect estimation of ground reaction force by using kinematic data has been considered due to the challenges that researchers face when using direct force measurement tools. Among indirect estimation methods, biomechanical and statistical models bring inaccuracy and uncertainty. As a flexible tool for nonlinear modeling and highly efficient, machine learning methods do not require prior knowledge of the model and create their model based on artificial neural network algorithms that find a strong relationship between input and target variables. For this reason, today the use of machine learning methods is of interest as a modern approach that simplifies modeling and data collection strategies for force estimation.


Main Subjects