Volume 27, Issue 1 (Spring--In Press 2026)
jrehab 2026, 27(1): 16-33 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Saburi L, Piri E, Jafarnezhadgero A. The Role of Artificial Intelligence in Human Gait Biomechanics: A Systematic Review. jrehab 2026; 27 (1) :16-33
URL: http://rehabilitationj.uswr.ac.ir/article-1-3711-en.html
URL: http://rehabilitationj.uswr.ac.ir/article-1-3711-en.html
1- Department of Sports Biomechanics, Faculty of Sports Sciences, Shahid Bahonar University of Kerman, Kerman, Iran.
2- Department of Sports Biomechanics, Faculty of Educational Sciences and Psychology, University of Mohaghegh Ardabili, Ardabil, Iran.
3- Department of Sports Biomechanics, Faculty of Educational Sciences and Psychology, University of Mohaghegh Ardabili, Ardabil, Iran. ,amiralijafarnezhad@gmail.com
2- Department of Sports Biomechanics, Faculty of Educational Sciences and Psychology, University of Mohaghegh Ardabili, Ardabil, Iran.
3- Department of Sports Biomechanics, Faculty of Educational Sciences and Psychology, University of Mohaghegh Ardabili, Ardabil, Iran. ,
Abstract: (1507 Views)
Objective This systematic review explores the transformative role of artificial intelligence (AI) in gait biomechanics, aiming to categorize current methodologies and evaluate their clinical efficacy.
Materials & Methods By searching major databases including PubMed, Scopus, and Web of Science, the study analyzed research published between 2020 and 2025. Following rigorous selection criteria focusing on experimental studies of human gait using AI, wearable sensors, or motion capture 13 high-quality articles were selected and assessed via the Downs and Black scale.
Results The findings highlight that wearable sensors are the predominant data collection tool (53.8%), followed by markerless systems and deep learning frameworks. The synthesis of evidence indicates that integrating wearable technology with machine learning (ML) models specifically stack and support vector regression (SVR) is exceptionally effective for classifying gait episodes, achieving a high mean sensitivity of 0.961 and a mean absolute error (MAE) of under 2.1% for primary biomechanical parameters. Advanced architectural combinations also showed significant promise; for instance, the ResNet101 and Naïve Bayes hybrid proved highly reliable for posture classification (sensitivity 0.87). Similarly, long short-term memory (LSTM) networks demonstrated remarkable accuracy in short-term gait path prediction within virtual reality settings, though long-term forecasting remains dependent on multi-modal data, such as eye-tracking. Finally, the application of smart insoles paired with random forest (RF) algorithms suggests substantial potential for identifying digital biomarkers in conditions like sarcopenia.
Conclusion In conclusion, while AI-driven approaches offer high precision in gait analysis, current research is often limited by small sample sizes. Future studies should prioritize larger cohorts to validate these models and enhance their clinical translation. This review confirms that AI, particularly when paired with high-fidelity sensor data, represents a powerful frontier for diagnosing and managing diverse neurological and musculoskeletal pathologies.
Materials & Methods By searching major databases including PubMed, Scopus, and Web of Science, the study analyzed research published between 2020 and 2025. Following rigorous selection criteria focusing on experimental studies of human gait using AI, wearable sensors, or motion capture 13 high-quality articles were selected and assessed via the Downs and Black scale.
Results The findings highlight that wearable sensors are the predominant data collection tool (53.8%), followed by markerless systems and deep learning frameworks. The synthesis of evidence indicates that integrating wearable technology with machine learning (ML) models specifically stack and support vector regression (SVR) is exceptionally effective for classifying gait episodes, achieving a high mean sensitivity of 0.961 and a mean absolute error (MAE) of under 2.1% for primary biomechanical parameters. Advanced architectural combinations also showed significant promise; for instance, the ResNet101 and Naïve Bayes hybrid proved highly reliable for posture classification (sensitivity 0.87). Similarly, long short-term memory (LSTM) networks demonstrated remarkable accuracy in short-term gait path prediction within virtual reality settings, though long-term forecasting remains dependent on multi-modal data, such as eye-tracking. Finally, the application of smart insoles paired with random forest (RF) algorithms suggests substantial potential for identifying digital biomarkers in conditions like sarcopenia.
Conclusion In conclusion, while AI-driven approaches offer high precision in gait analysis, current research is often limited by small sample sizes. Future studies should prioritize larger cohorts to validate these models and enhance their clinical translation. This review confirms that AI, particularly when paired with high-fidelity sensor data, represents a powerful frontier for diagnosing and managing diverse neurological and musculoskeletal pathologies.
Type of Study: Systematic Review |
Subject:
Sport Pathology and Corrective Movements
Received: 3/12/2025 | Accepted: 2/02/2026 | Published: 1/04/2026
Received: 3/12/2025 | Accepted: 2/02/2026 | Published: 1/04/2026
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
Copyright © The Author(s);
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC-By-NC), which permits use, distribution, and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
Contact Information
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC-By-NC), which permits use, distribution, and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
Contact Information
