Development of a shoulder muscle feedback controller for human body models.

Introduction: State-of-the-art finite element human body models (FE HBMs) with active muscle controllers can predict occupant kinematics during braking and steering, which are typical pre-crash interventions aiming at avoiding crashes. Information about the pre-crash occupant kinematics can be used in the design of systems that influence the occupant position in the pre-crash phase and the interaction between the occupant and the restraints in both the pre- and in-crash phases. For driver HBMs, active shoulder muscles are required to reproduce the load between the steering wheel and the torso. The shoulder is the most freely moving joint in the body, and the stability of the shoulder complex depends on muscle activity. Thus, intermuscular load sharing cannot be determined solely from the geometrical location of the muscle because other muscles co-contract to maintain stability during the movement. The aims of this study were to implement a new controller, which introduces load sharing based on physical tests with volunteers, into the shoulder of an FE-HBM and to compare its performance with that of volunteers subjected to dynamic elbow loading. Methods: A new shoulder muscle controller for use in FE-HBMs was developed, including directionally dependent intermuscular load sharing based on recorded muscle activity from volunteers. The controller performance was evaluated by simulating a volunteer experiment, exposing the subjects to dynamic loading of their elbow in eight directions. Results: Elbow kinematics were compared between the model and volunteers. A sensitivity study was also performed to evaluate the controller gains. The model successfully predicted peak elbow displacements for all loading directions. Discussion: One limitation in the current study was the use of a submodel and a simplified experimental setup. In a braking or steering maneuver, head and torso inertia would introduce forces to the shoulder, instead of forces introduced in the elbow as in this study. Because these two scenarios are mechanically similar, a simplified approach was used instead, as this allowed for an experiment where the force magnitude and direction could be easily controlled. Hence, the developed shoulder muscle controller is ready to be implemented and evaluated in a full-body active FE-HBM exposed to driver maneuvers.