This paper describes a method of redundancy resolution that combines closed form inverse kinematics with multicriteria optimization to form a method able to coordinate the motion of a hyper-redundant serial robot at deterministic speeds greater than 100 cycles per second on a personal computer. The work includes a listing of over 30 task-based performance criteria derived from physical models of the robot and presents formulations for several of them. It discusses a motion coordination method that explicitly generates motion options and evaluates them based on any number of performance criteria. The highest-ranking option becomes the next motion command for the robot's servo controllers. Finally, the paper presents a simulation of motion coordination for a hyper-redundant serial robot with 21 degrees of freedom.
展开▼