Biomechanical behavior of lower limb joints based on Time-series k- means clustering Algorithm during Gait

Abstract Given the inherent essence of gait, there is potential to analyze and categorize this data using artificial intelligence. Aim of this study Investigate the mechanical properties of lower limb articulations while walking through the application of artificial intelligence methodologies. Twenty-eight young individuals were part of a study. They completed a task walking 10 meters on a path. The task was done twice, once at their own speed and once at 100 beats per minute. Data points were categorized using a machine learning technique called time-series k-means clustering. The movements of hip, knee, and ankle joints were compared across different clusters. Statistical analysis was used to find any differences among the groups. Results of the study shows that angular velocity and joint moment influence subject distribution in clusters. Cluster 3 with the smallest population has the highest mechanical power. Mechanical power of ankle and hip joints affects clustering pattern. Second group has few participants with random power allocation among joints. Knee joint is the primary determinant in clustering process. Hip joint has a minor role in power generation and absorption at preferred speed. Knee and ankle joints are crucial in clustering outcomes in all scenarios. The findings indicate notable disparities between the production and absorption stages of walking in situations involving preferred and constant speeds, with the exception of the hip joint.