An analysis is made of the experimental results of a minimum-time trajectory control scheme for a two-link flexible robot. An offline optimization routine determines a minimum-time, straight-line tip trajectory which stays within the torque constraints of the motors and ends with no vibrational transients. An efficient finite-element model is used in the optimization to approximate the flexible arm dynamics. The control strategy described is used to determine the feedback gains for the position, velocity, and strain gage signals from a quadratic cost criterion based on the finite-element model linearized about the straight-line tip trajectory. These feedback signals are added to the modeled torque values obtained from the optimization routine and used to control the robot arm actuators. The results indicate that this combination of model-based and error-driven control strategies, achieves a closer tracking of the desired trajectory and a better handling of modeling errors than either strategy alone.
展开▼
