ABSTRACT
Due to variability in human anthropometry, no two people will perform a task in exactly the same way—different people will adapt different postures to perform the same reach task, and even for the same person and reach task, postures will vary with time. It is therefore important to consider uncertainty in methods of posture prediction. Currently, techniques in digital human modeling applications primarily employ deterministic methods which do not properly account for variability. An alternative to deterministic methods is probabilistic/reliability design. This study presents a sensitivity approach to gain insights into how uncertainty affects reach tasks. Sensitivity levels are found to determine the importance of each joint to the final posture. A digital human upper body model with 30 degrees of freedom (DOFs) is introduced to demonstrate the sensitivity approach for reach tasks using both arms. Thirty-six different reach tasks (eighteen for male and eighteen for female) are used to compare the sensitivities due to joint angle and link length uncertainty. The results show that the importance of each joint and link length is dependent on the nature of the reach task; sensitivities for joint angles and link lengths are different for each reach task.
