Adaptive control techniques and feed forward compensation of periodic disturbances in industrial manipulators

This paper considers the compensation for periodic torque disturbances, characteristic of electrical permanent magnet motors. These disturbances are sinusoidal with known but rapidly varying frequencies. The compensation is obtained as a two stage adaptive controller. The final compensator is obtained as a pure predictive feed forward compensation that includes an internal model of the disturbance. The disturbance's internal model identification is obtained applying a closed loop, adaptive feed forward control algorithm. The stability and robustness properties of the overall adaptive compensation technique are validated by experimental tests on industrial manipulators.