Feedforward and of a Flexible

Feedforward and feedback control strategies are applied to a servo- driven flexible structure and studied in time- domain computer simulations. Simulation results are compared to experimental runs on a flexible arm apparatus. The most accurate and stable responses of the arm are obtained by adding corrective terms to the command signals when position feedback is taken only from the servo. Full conditioning of the command signal, which involves compen- sating for the dynamics of both the servo and the arm, produced the best response. Simu- lations and experiments show that positional and accelerometric feedback, as combined with a proportionalderivative element acting on the position error, is a viable solution in the less stable situation when feedback is being taken only from the manipulator. DC servomotor driving a slender aluminum beam. The purpose of the experiments was to identify simple and effective control strat- egies to deal with the motion errors that oc- cur when a servomotor is driving a very flex- ible structure. Timedomain simulations were developed to model the behavior of the servo and the arm, and were compared to the ex- perimental results. Since robot controllers have extensive computational overhead, em- phasis was placed on using the simplest lin- ear models of the servo and the arm that still predicted the motions accurately. Two kinds of control strategy were inves- tigated. In the first, the motion command signal was used to compensate for the dy- namics of the positional servo and the ann. No position feedback was taken from the ann tip. This is often refed to as feedforward control (5), 161. In the second approach, compensators were used in a feedback con- figuration, with the arm tip location fed back to the motor controller and the motor oper- ated as a velocity servo. There are other, more sophisticated, control strategies, but they were not considered here. In addition, note that the feedforward strategy has its limitations. The controller cannot know if the arm has reached the target, and there is no allowance for disturbance rejection.