Time-scaling of trajectories for point-to-point robotic tasks

This paper addresses the problem of robotic tasks consisting in time-optimal point-to-point motion under limited torque input. This class of tasks can be performed by the classical approach of trajectory tracking control using a minimum-time trajectory. However, the tracking of the time-optimal trajectory leaves no control faculty to compensate for tracking errors entailed by model uncertainties. In order to overcome the performance errors, in this paper is introduced an algorithm that search for on-line time-scaling of the time-optimal point-to-point trajectory. A numerical example for a two degrees-of-freedom arm is provided, illustrating the tracking accuracy of the proposed method.