Model-Predictive Vibration Reduction of a Suspended Cable-Driven Parallel Robot

This letter focuses on the vibration reduction for Cable-Driven Parallel Robots. A Non-linear Model Predictive Controller (NMPC) is designed and implemented to adapt the control inputs sent to the robot actuators. This NMPC is based on an elasto-dynamic model of the robot, which is discretized and included in a constrained optimal control problem. Two experiments are conducted and the performance of the NMPC is compared with a Computed Torque Controller and a NMPC with an inelastic model. The first one is a trajectory following task and shows that the proposed controller significantly reduces the amplitude of the vibrations of the Moving-Platform along the trajectory and the second one focuses on a sudden braking situation and shows that the proposed control scheme manages to divide the amplitude of the post-braking oscillations by factors ranging from one and a half to twenty-two, demonstrating the interest of this controller for perturbation rejection.