Human Grasp Modeling

Digital human modeling for ergonomics applications is the focus of my research in the Center for Ergonomics. The research I direct in the Human Motion Simulation Laboratory focuses on the development of algorithms and models for accurate prediction of human postures and motions in tasksituations relevant to product and workspace design. We emphasize algorithms that can be implemented in any human figure model, rather than limiting ourselves to the capabilities of one model. We are fortunate to have the developers of the two most popular software systems for industrial ergonomics (Dassault Systemes and Siemens PLM) engaged in our laboratory as technology partners. Among other topics, we're currently working on improving grasp simulation. The literature on grasp is vast, including contributions from motor control and robotics as well as ergonomics and human modeling. Yet, ergonomics using human models spend a large amount of time working to simulate grasp, often with poor results. As part of ongoing research in the Human Motion Simulation Lab, Ph.D. candidate Wei Zhou and I developed a data-based method for predicting grasp motions. The hand motions are parameterized on target size and grasp type, so that they're readily configurable for different objects. We integrated this kinematic prediction with collision sensing to enable automated grasp of objects with arbitrary geometry.