Artificial Intelligence in Sports Injury Prediction, Prevention, and Management: A Systematic Literature Review and Recommendations for Future Research

The rapid development of artificial intelligence (AI) and related technologies (machine learning, deep learning, computer vision, wearable sensors and IoT) has revolutionized sports medicine, particularly in the prediction, prevention, diagnosis, rehabilitation and management of athletic injuries. Despite the increasing number of studies applying AI to sports injury, there is no systematic and comprehensive review that critically analyzes the current mechanisms, performance metrics and remaining challenges. This paper presents a systematic literature review (SLR) of ۲۰ high-quality studies published between ۲۰۲۳ and ۲۰۲۵. The selected mechanisms are classified into four main categories: (i) classical and ensemble machine learning for injury risk prediction, (ii) deep learning and computer vision for biomechanical analysis and pose estimation, (iii) hybrid AI-IoT/wearable systems, and (iv) AI applications in diagnosis, rehabilitation and mental health. A detailed comparison using key performance metrics (accuracy, AUC, sensitivity, specificity, latency, explainability, data requirements, and clinical applicability) is provided. The results show that ensemble methods (Random Forest, XGBoost) and deep learning models (CNN, LSTM, RNN) consistently achieve ۸۵–۹۷ % accuracy in injury prediction and classification tasks. However, critical challenges remain, including lack of standardized open datasets, limited model explainability (black-box problem), ethical and privacy concerns, and insufficient prospective clinical validation. This review offers clear guidelines and recommendations for future research.

کلیدواژه ها:

Artificial Intelligence ، Machine Learning ، Deep Learning ، Sports Injury Prediction ، Injury Prevention ، Wearable Sensors ، IoT ، Biomechanics ، Systematic Literature Review