Model Predictive Control for Cable-Driven Parallel Robots Using Moving Horizon State Estimation

Taking advantage of fast motion, strong load capacity, and large working space of cable-driven parallel robots (CDPRs), it is newly developed and applied as the capture system for reusable rockets to improve the success of recovery. However, it is a challenge to precisely control the end-effector of CDPRs to catch the landing rocket under the disturbances of the impinging exhaust jet and plume distribution and undergo the inherent unidirectional force characteristic of CDPRs. To deal with this challenge, the combination of the model predictive control (MPC) algorithm and the moving horizon estimation (MHE) method is employed in the capturing process of CDPRs. These two methods are based on the optimization method. The pseudo-drag problem due to the unidirectional force characteristic is solved by MPC, and the disturbances such as process noise and measurement noise on the state feedback loop can be filtered by MHE method. The precision and performance of the combination control scheme are verified and compared by some benchmark simulations.