AI-Based Analysis of BMI Impact on Gait Dynamics Across Speeds

Gait analysis plays a critical role in understanding human biomechanics and its relationship to health, offering insights into mobility and gait-related disorders. This study investigates how Body Mass Index (BMI) categories (normal, overweight, and obese) influence gait dynamics across three walking speeds ($0.5 \mathrm{~m} / \mathrm{s}, 0.75 \mathrm{~m} / \mathrm{s}$, and $1.0 \mathrm{~m} / \mathrm{s}$) using data collected from a split-belt treadmill with integrated force plates. Sixty healthy participants were categorized into BMI groups according to WHO standards. Key spatiotemporal gait parameters, including ground reaction forces, joint moments, and center of pressure trajectories, were analyzed to identify BMI-specific biomechanical differences. Machine learning models, including Random Forest and Neural Networks, achieved high accuracy in classifying gait speeds within each BMI group, with Random Forest achieving up to $99.10 \%$ accuracy in the Overweight group. Results revealed that increasing BMI led to distinct biomechanical adaptations, such as higher joint loads and variability in obese individuals. These findings demonstrate the potential of combining BMI-specific gait analysis with machine learning to enhance personalized rehabilitation and clinical assessments.