A space state approach to flexible robotic joint control

After reviewing the control schemes commonly adopted for the control of an industrial robotic joint, the authors propose a state space approach to this problem which includes the disturbance suppression on both the motor and the arm sides. The use of low-resolution encoders is considered, and the usefulness of a Kalman filter for motion quantities and disturbance estimation is pointed out. Two different estimator schemes are proposed and compared on the basis of computational load and system performance. The effectiveness of the proposed solutions has been tested on an experimental system, realized with a low-stiffness robotic joint, driven by a DC motor, and a digital signal processor (DSP)-based control system. The results of the tests are presented, along with the discussion of the merits and limits of this approach.