Design and Evaluation of a Machine Learning-Based Biomechanical ‎Feedback System for Correcting Landing Pattern in Student ‎Athletes

Introduction: Improper landing patterns are among the most important risk factors for lower ‎extremity injuries in student athletes. Traditional methods for correcting landing patterns are ‎based on visual feedback from coaches and lack sufficient accuracy and personalization. This ‎study aims to design and evaluate a machine learning-based biomechanical feedback system for ‎correcting landing patterns in student athletes.Methodology: Sixty student athletes (۳۰ females, ۳۰ males) with a mean age of ۱۲.۴±۱.۶ years ‎were randomly divided into experimental and control groups. Kinematic landing data were ‎collected using inertial sensors. A random forest model was designed to classify landing ‎patterns. The experimental group received machine learning-based feedback through a user ‎interface for ۸ weeks. Data were analyzed using paired t-tests and analysis of covariance.Findings: The machine learning model identified unsafe landing patterns with ۹۲.۲% accuracy, ‎‎۸۹.۵% sensitivity, and ۹۲.۶% specificity. After the intervention, the experimental group showed ‎significant improvements in knee flexion angle (from ۳۱.۴±۵.۲ to ۳۸.۷±۷.۸ degrees; P<۰.۰۰۱), ‎reduction in knee valgus angle (from ۱۴.۲±۶.۳ to ۹.۸±۴.۴ degrees; P<۰.۰۰۱), and reduction in ‎vertical ground reaction force (from ۲.۷±۰.۴ to ۲.۱±۰.۳ times body weight; P<۰.۰۰۱) compared ‎to the control group. Effect sizes for the main variables ranged from ۰.۷۲ to ۰.۸۵, and the injury ‎risk score in the experimental group decreased by ۴۶.۲%‎.Conclusion: The machine learning-based biomechanical feedback system, by providing accurate ‎and personalized feedback, can serve as an effective tool for correcting landing patterns and ‎reducing injury risk factors in student athletes, and has high potential for improving the quality ‎of motor learning and preventing sports injuries in young populations.

کلیدواژه ها:

Biomechanical Feedback ، Machine Learning ، Landing Pattern Correction ، Student Athletes ، ‎Injury Prevention.