Faculty Mentor(s)

Dr. Brooke Odle, Engineering

Document Type


Event Date



Manual patient-handling tasks are associated with lower-back pain and injury. Computational musculoskeletal models may determine forces on the low back, trunk muscle activation during these tasks, and provide insight on how these may result in injury. There are no publicly available models for these tasks. To determine the appropriateness of the OpenSim Full-Body Lifting Model for these tasks, this project calibrated the model with biomechanical data of simulated maneuvers performed by a test subject. Thirty-nine reflective markers were placed bilaterally on bony prominences (forehead, back of head, clavicle, sternum, C7, T10, epicondyle of humerus, styloid process, metacarpal, anterior and posterior iliac crests, calcaneus, toe, upper arm, forearm, thigh, and shank). Ten electromyography (EMG) sensors were attached to the bilateral external obliques, rectus abdominis, thoracic and lumbar erector spinae, and rectus femoris. The subject performed three maneuvers 15 times while standing on two force plates: a twisting motion at the hips, a one arm raise, and a shallow squat. Kinematic data were collected at 100Hz, and EMG and force plate data at 1000Hz. The kinematic and kinetic data were applied to the model to scale and perform inverse kinematics and dynamics. It is recommended that the maximum marker errors be between 2 and 4 centimeters for inverse kinematics and about 2 centimeters for scaling. However, the acceptable error may vary by application. In our trials the maximum errors for inverse kinematics varied between 1 and 10 centimeters. During a lifting trial, the left forehead marker was the only marker with high error. Since that marker is mainly important for scaling, a high error in inverse kinematics shouldn’t affect the simulation predictions; so our calibration procedure ignores high errors from non-essential markers. Our calibration procedure will provide a framework for evaluating simulated results and the feasibility of the model for patient-handling tasks.


This material is funded by the Michigan Space Grant Consortium, NASA grant #NNX15AJ20H, and the Clare Boothe Luce Research Scholars Program of the Henry Luce Foundation.