Manipulators have similar structure with human upper extremity, and there are many possibilities to utilize them for various works instead of humans. Hence, human arm movements have been widely studies, which analyze sophis-ticated capability of human arm control system and attempt to improve ability of manipulators by transferring the results of such analysis.
It is reported that human arm trajectories have particular properties such that most of its path lies along a rather straight or simple curved line, and a tangential velocity of hand is bell shaped. In this paper, human arm trajectories are analyzed in Cartesian space and joint angle space, and performance criterion is proposed to formulate trajectories for demonstrating human arm movements. Moreover, difficulty of solving nonlinear optimization problem which disturbs manipulators from application of trajectory planning is overcome by developing an algorithm of near-optimal trajectory formation based on approximated differential equations. Besides, it is confirmed that optimal trajectory with high accuracy is stably formulated by utilizing the algorithm
Formulated trajectories according to a proposed trajectory planning strategy are compared with observed trajectories obtained by experiment which is conducted under unrestricted conditions. The results show that the formulated trajectories well demonstrate human arm trajectories.