Rapid Analysis of Human Musculoskeletal Dynamics Combining Markerless Motion Capture, Extended Kalman Filtering and Semi-Recursive Method

Human musculoskeletal modeling provides a quantitative analysis method for obtaining joint angles, torques, and muscle contraction patterns. However, its kinematic inputs rely on infrared motion tracking, difficult to apply to out-of-lab tracking practice. Conventional musculoskeletal dynamics modeling using differential-algebraic equations introduces redundant degrees of freedom (DOFs). This study proposes a simulation framework combining markerless motion tracking, inverse dynamics simulation, and muscle redundancy calculations. The sagittal gait kinematics measured by markerless motion tracking are filtered by extended Kalman filtering. Using the obtained data as kinematic inputs, a multibody dynamics model is further developed based on the semi-recursive description. By this means, the computational speed can be fastened by changing its governing equations to ordinary differential ones without constraints. Numerical results validate the efficiency and accuracy of the proposed approach, highlighting its potential use in clinical and sports engineering scenarios.