Real-Time Dynamics Simulation of the Upper Limb Musculoskeletal Multibody System Based on Semi-Recursive Method*

Musculoskeletal biomechanics simulation is crucial for evaluating the risk of muscle fatigue and joint injuries among manual handling workers caused by high-intensity repetitive upper limb operations. This study proposed an inverse dynamics simulation framework for the upper limb musculoskeletal system based on a semi-recursive approach. The upper limb kinematics data are measured by inertial measurement units (IMUs) and filtered by Extended Kalman Filter (EKF). A musculoskeletal model of the human upper limb is further established for estimating the joint torques and muscle recruitment patterns through inverse dynamics methods. The semi-recursive approach is further adopted to simplify the multibody dynamic formulations through velocity transformation. Numerical simulations are validated against conventional modeling methods and surface electromyography (sEMG) data. The results demonstrate that the proposed method enables real-time estimation of muscle forces and joint loads, providing theoretical foundations for exoskeleton design and personalized adaptation.