Abstract
We have already formulated a human three-joint arm's optimal control model with a freezing-like mechanism in its hand joint and clarified the model's effectiveness in reproducing two-point reaching movement characteristics. This research formulates a new optimal control model for reproducing human arm's obstacle-avoiding movements through incorporating two kinds of obstacle-avoiding conditions into the three-joint arm's optimal control model and examines the model's performance. The first condition is that the hand path is tangent at an avoiding time to the avoiding line that is parallel to the line connecting the start and final points. The second condition is that the tangent point, i.e., the avoinding point, involves the position of the obstacle deeply. Consequently, the following results are obtained: (1) the proposed model can generate obstacle-avoiding movements similar to measured ones under the two avoiding conditions and is superior in reproducibility to the previous model; (2) there exists the optimal avoiding time to minimize the model's criterion function; (3) the model's optimal obstacle-avoiding trajectories based on torque-change minimization agrees better with the measured ones than those based on energy minimization. These results suggest that the model's avoiding conditions can be valid in reproducing human arm's obstacle-avoiding movements and that the human arm's obstacle-avoiding mechanism can function based on torque-change minimization.
