| The proposed research deals with digital human modeling and simulation. Digital humans are avatars that are digitally created, have the appearance of human-like motion and behavior, and are used to simulate human motion and performance. Digital humans have become a fundamental cornerstone of engineering analysis towards achieving a higher level of digital prototyping. Our proposed work deals with predicting static and kinematic motions of digital humans, in the most realistic manner possible, to simulate their existence in a virtual world and to enable them to test and experience products that are only defined in the digital world, thus reducing the time and cost associated with prototyping. To achieve this goal, we have started with a significantly larger number of degree-of-freedom model than typically used by researchers. We have then introduced a unique task-based approach to posture prediction as a postulate for why people assume specific postures. This postulate led to an optimization-based approach, where human performance measures are quantified as functions of variables that evaluate to real numbers, thus can be implemented in a multi-objective optimization algorithm for arriving at “the best” posture. Path trajectories followed by humans in space to execute a task were also addressed. The concept of admissible kinematically-smooth trajectories was created to characterize a path that does not admit switching of inverse kinematic solutions during motion, therefore produce realistic smooth motion…a concept that is true for humans but not for robotic motions. We have also investigated the prediction of joint variables as vector functions of time to predict how human upper body (including the upper extremities) behaves as the hand moves between any two points in space. The end result is an optimization-based method using human performance measures such as discomfort and smoothness in combination with minimum jerk model for calculating joint path trajectories that look and feel most natural. Long term goals of this research are to enable autonomous behavior and realistic motion of digital humans, with the ultimate goal of reducing or eliminating the use of prototypes in the design cycle. |
