An online learning procedure for feedback linearization control without torque measurements

By exploiting an a priori estimate of the dynamic model of a manipulator, it is possible to command joint torques which ideally realize a Feedback Linearization (FL) controller. The exact cancellation may nevertheless not be achieved due to model uncertainties and possible errors in the estimation of the dynamic coefficients. In this work, an online learning scheme for control based on FL is presented. By reading joint positions and joint velocities information only (without the use of any torque measurement), we are able to learn those model uncertain- ties and thus achieve perfect FL control. Simulations results on the popular KUKA LWR iiwa robot are reported to show the quality of the proposed approach.