An algorithm is presented for the on-line generation of minimum-time trajectories for robot manipulators. The algorithm is designed for intelligent robots with advanced on-board sensory equipment which can provide the position and orientation of the end-effector. Planning is performed in the configuration (joint) space by the use of optimised combined polynomial splines, along with a search technique to identify the best minimum-time trajectory. The method proposed considers all physical and dynamical limitations inherent in the manipulator design, in addition to any geometric path constraints. Meeting the demands of the heavy computations involved lead to a distributed formulation on a multiprocessor system, for which an intelligent control unit has been created to supervise its proper and practical implementation. Simulation results of a proposed case study are presented for a PUMA 560 robot manipulator.

A method for trajectory control in the joint space is presented. An acceleration profile is proposed for each segment of the trajectory. After a twofold integration a position trajectory is obtained with advantageous characteristics. The position trajectory is completely dynamically balanced; it exhibits continuity up to the third derivative of the position. This way, minimum requirements are imposed on the actuators. The technique delivers predictable results since the trajectory deviates only slightly from a straight line connection between successive joint coordinates. Very limited computational effort is required.