A cartesian-based adaptive tracking controller for a SCARA robot

This paper describes the real-time implementation of a cartesian-based controller applied to a 2 DOF direct-drive SCARA robot. This control scheme uses directly path shapes described in cartesian coordinates and therefore avoids the trajectory conversion in joint-based coordinates. This results in a greater precision and reduces drastically the off-line calculations. However, the cartesian-based controller must perform many on-line computations because of the kinematics and other transformations. Moreover, the authors´ controller is adaptive and takes into account the full dynamics of the robot. Nevertheless, the authors´ cartesian-based adaptive controller is implemented on a single-chip digital signal processor (DSP96002) with minimal hardware. The performance of the controller is illustrated by experimental results showing the tracking of a triangular trajectory. Results obtained with a simple PD controller are also shown for comparison