The very essence of robotic manipulation tasks lies in the interaction of the manipulator with objects in their workpiece or environment. Tasks are divided largely into the three ones : non-contact task, contact task and impact phenomenon as a transition process from non-contact task to contact task. In order to effectively control a series of the three tasks, position/impact/force control is needed.
Redundant robot manipulators have a lot of advantages compared with nonredundant manipulators and provide increased flexibility and dexiterity for the executation of complex tasks. The redundancy can be effectively used to avoid singularities, optimize joint torques, and reduce the effects of impact while performing the desired end-effector task.
In this thesis, a strategy for position/impact/force control of robot manipulators with kinematic redundancy is proposed. For the proposed strategy, effective control algorithm of arm configuration and performance measures which express the optimum configuration for the given task are required.
First of all, we propose an effective real-time dynamic control algorithm for redundant manipulators which optimizes arbitrary performance measure using internal self-motion capability while tracking a desired end effector trajectory and shows low computational burden compared with the conventional method such as the method of Hsu et al. Second, we investigate the optimum configurations for reducing impact effects and improving contact task. Finally, through switching the performance measure, we show that optimum configrration for each task is obtained when manipulator faces a series of the tasks.
The effectiveness of the proposed strategy is illustrated with computer simulation results.
