Hamstring Injury Prediction Using AI and Biomechanics

Problem addressed:
Hamstring injuries are among the most frequent and costly injuries in football, often occurring during high-speed running, sprinting, and rapid acceleration phases. These injuries are strongly associated with excessive mechanical loading, muscle fatigue, and poor recovery. Current monitoring approaches rely mainly on external load indicators (e.g., GPS metrics), which fail to capture internal muscle stresses and strain accumulation, limiting their predictive accuracy.

Proposed solution:
This project aims to develop a hybrid predictive framework integrating:

• Biomechanical modeling of lower limb dynamics (muscle forces, joint kinematics)

• Muscle fatigue and strain accumulation models inspired by damage mechanics

• Physics-informed neural networks (PINNs)

• Surrogate AI models for fast and real-time prediction

• Integration of wearable data (GPS, IMU) for player-specific assessment

The objective is to predict injury risk before critical muscle damage occurs and provide actionable recommendations for load management.

Key applications:

• Injury prevention strategies in professional football clubs

• Player-specific training load optimization

• Early warning systems for medical and coaching staff

• Return-to-play decision support after hamstring injury

Expected impact (economic and performance):

• Reduction in hamstring injury incidence and recurrence

• Improved player availability and performance

• Reduced medical and rehabilitation costs

• Enhanced decision-making using data-driven insights

Collaboration sought:

• European research groups in biomechanics, sports science, and AI

• Football clubs, federations, and sports analytics companies

• MSCA Postdoctoral candidates in biomechanics and intelligent systems

What I offer:

• Expertise in computational mechanics and fatigue modeling

• Strong background in physics-based and AI-driven approaches

• Focus on translating engineering models into practical sports applications