Browse Publications Technical Papers 2009-01-0921

Posture Prediction with External Loads – A Pilot Study 2009-01-0921

As the need for more advanced human modeling tools has grown, so has the focus on research and development with posture-prediction capabilities for the design and analysis of products, and for the study of human behavior. Virtual humans have grown from digital mannequins with limited fidelity, to realistic avatars with predictive capabilities. Now, one of the frontiers with posture prediction is the incorporation of external loads and joint torques. Although advancements have been made with dynamic motion prediction, relatively little work has been conducted with external load-based posture prediction. Drawing on past success with optimization-based kinematic posture prediction implemented with the virtual human Santos, we present a new method for considering external loads. A pilot study is conducted whereby equations for static equilibrium are incorporated in the optimization formulation. Consequently, torque (as well as joint angles) is determined for each degree of freedom, and is incorporated in human performance measures that serve as objective functions in the optimization formulation. The intent is to test the feasibility of extending the formulation for kinematic posture prediction. Different external loads are applied to the right and left hands respectively, while the same target points are provided for these different lead cases on both hands. The results for the previous work (kinematic formulation) and the new formulation are compared. The predicted postures are evaluated quantitatively in terms of numerical output, and subjectively in terms of visible postures. In general, the pilot study was successful; the predicted postures were reasonable. Including torque provided more realistic predicted postures and sets the stage for new discomfort models as well as consideration of reaction forces.


Subscribers can view annotate, and download all of SAE's content. Learn More »


Members save up to 17% off list price.
Login to see discount.