Nonlinear MPC for free-flying space manipulator

The growing number of space debris orbiting earth threatens access to space. Human intervention has become necessary to avoid the “Kessler Syndrome” and maintain Earth’s orbit accessible. A manipulator mounted on satellite could be used to capture and de-orbit critical debris. These space robots could also be used for other orbital applications such as maintenance and assembly. However, the space environment and dynamic coupling between the base and manipulator render their modelling and control highly challenging. This research aims at developing a new control method for precise guidance of the end-effector to capture a tumbling target. In this thesis, a nonlinear model predictive controller (NMPC) is used to control the end- effector of a free-flying space robot in the task space with the presence of parametric uncertainties. In the free-flying mode, the base actuators are used for the control task. Before developing the theory behind NPMC, the kinematics and dynamics of the nonlinear model are obtained using the Decoupled Natural Orthogonal Complement (DeNOC) approach. The validation of the controller being almost impossible to do on Earth, a high-fidelity model is developed. The control technique is validated for three different scenarios using a space robot mounted with a six degrees of freedom (DoF) manipulator. The simulations are performed in the Matlab/Simulink environment. In the first scenario, the performance of the NMPC for the trajectory tracking of a circle is analyzed. The second scenario tests the robustness in the presence of parametric uncertainties. The same trajectory is considered but with uncertainties on the base’s parameters. Finally, the last scenario tests the controller for a different trajectory than the one used for tunning.