Guidance and Control in Space Debris Removal Missions via Adaptive Nonlinear Model Predictive Control

Space debris orbiting around the Earth is becoming a major problem that could impair the future of space exploration. This problem was formulated the first time in the 1960 by the scientist Willy Ley. Then, in the 1978 the scientist Donald J. Kessler formulate a possible dystopian scenario that takes the name of Kessler syndrome. Among the different approaches to this problem that have been proposed in recent years, this thesis focuses on a possible solution exploiting an automated method. One important challenge of this approach is that when the space debris is collected by the spacecraft (S/C), its unknown mass will affect the dynamics of the whole system. In fact, this could be the case where the information on the S/C dynamics alone is not sufficient to successfully accomplish the different tasks of a space mission. The aim of this study is the design of a control system capable of handling situations where not all the parameters are known a priori or change in time. These parameters are needed in order to accomplish the Guidance Navigation and Control (GNC) problem of the debris removal mission. For this purpose, a control strategy based on an adaptive version of the Nonlinear Model Predictive Control (NMPC) is accounted for. This control approach is implemented ad hoc in the debris removal space mission, where the unknown debris mass, treated as a parameter in the control problem, is estimated with two different methods (Recursive Average and Extended Kalman Filter (EKF)) in order to obtain an optimal trajectory needed for the debris successfully removal. Simulations shows that the designed control system is able to accomplish optimally this task. The solutions coming from the two different estimation methods are compared. It will see that the obtained results are similar, but with the EKF case it is possible to increase the state measurements (in the case of noises coming from sensors). If the simple NMPC approach is employed without an estimation process, the obtained results do not diverge significantly from the optimal ones. This last result shows that the robustness of this control method is high. Anyway, the mass estimation could be useful in order to increase precision and performances. The Scope of this work is reached and the adaptive algorithm can be also extended to other space missions and/or to other engineering fields.