Velocity estimation by digital filtering of position data to process adaptive control laws for robotics manipulators

This paper deals with the study of some estimation methods of velocity using only position data, with the particular aim of processing adaptive control laws for robotic manipulators so that the harmonic content of the mechanical torques developed at the joints be reduced and the position errors be sufficiently small when trajectory tracking is the specific task. Moving average filters and least square filters are, first of all, reviewed and then a new method is discussed which is suitable for velocity estimation in closed loop control systems specifically designed for the tracking of trajectories. Experiments, carried out on a closed loop control system consisting of a planar manipulator and an adaptive controller which uses velocity data obtained by applying above methods, are described.