Active Damping of Parallel Robots Driven by Elastic Cables using On-Off Actuators through Model Predictive Control Allocation

摘要

This work studies the vibration rejection on elastic cable-driven parallel robots (CDPRs). Additional cold-gas thrusters are embedded on the robot in order to improve the rejection bandwidth. Such Unilateral Force Generators (UFGs) work as on-off actuators. Under the framework of optimal control, a Model Predictive Control (MPC) is designed to compute the control law and allocate the control signals to the available actuators by assigning their binary on-off states, thus forming a Mixed Integer Quadratic Programming (MIQP)-based MPC controller. Simulations highlight the benefits of the proposed predictive approach that yields a better rejection, fuel efficiency, and a reduced switching between ON and OFF states with respect to previous approaches.