Abstract
Motion of robot manipulators can be divided generally into the following three categories: i) fine motion which requires high positioning accuracy in all directions, but makes light of the moving speed (ex. assembling) ii) gross motion which tolerates rough positioning accuracy in all directions, only requires high speed movement (ex. material handling) iii) partially fine motion which requires high speed movement in sole directions and high positioning accuracy in other directions (ex. grinding). In this paper we propose several control schemes which are suitable for realizing the motion of category ii) or iii) by extending the method devised for the motion of category i) by Takegaki and Arimoto. The basic idea to derive these schemes is to supply the manipulator system with an artificial potential function and dissipation function according to the substance of each task. In the paper, two potential functions, ‘pyramid’ potential and ‘cone’ potential are presented in order to realize the motion of categories i) and ii). Moreover, we introduce a local coorinate system on the trajectory of end effector to be traced and define a particular potential function for the motion of categories i) and iii). The proposed control schemes strictly assure the asymptotic stability of the target point. It is also possible to incoorporate various sensory devices into them. The effectiveness of the proposed control schemes is verified by several experimental results.
