Task driven feedback control of robot arms: A step toward intelligent control

Abstract

The process of connecting task descriptions originating from machine intelligence planning programs to the mechanization of feedback control of robot arms is analyzed. It is shown in this paper that control theories and practices can be extended to a higher level where feedback control of robot arms directly can respond to work space task commands provided that the work space task as a command is given in the form of a closed function of time. A general mathematical procedure using tools from differential geometry is introduced for synthesizing task space motion planning so that the planned motion can be used as a direct input to the robot arm feedback control system to achieve desired robot hand motion. By definition, "intelligent control" is being manifested through robot performance in the task space relative to task space commands. Thus, the capability of implementing feedback control of robot arms directly driven by appropriate task descriptions in the workspace as commands is a step toward intelligent control.