Spacecraft Autonomous Guidance and Control

This thesis aims to develop an autonomous guidance and control for Earth monitoring missions, based on the Nonlinear Model Predictive Control (NMPC) approach, alternative to the standard planning designed on ground. As case study, the scenario in which a spacecraft (S/C) has to overfly a given point located on the Earth surface, when the on-board software is triggered by an alarm, is considered. The satellite must guarantee a fast revisit and persistent monitoring of the event and, when the alert is over, the return to the initial orbit. As Earth observation mission, the ESA Sentinel-2 one is studied. Firstly, the strategies for the autonomous mission planning are developed. Their goal is to produce feasible trajectories for guiding the S/C exactly on the coordinate in which the alert is occured. Then, the NMPC is designed in order to generate a control input which forces the S/C to follow the reference trajectories, finding a suitable trade-off between the time to perform the maneuvers and the propellant consumption. Constraints on the maximum thrust supplied by the engines and on the S/C orbital dynamics are taken into account explicitly in the problem. The obtained results are compared with the ideal strategy, which consists in instantaneous and impulsive maneuvers applied at the orbital node, and with the standard mission planning, in which the ideal situation is approximated with a consequent efficiency degradation. The results show that the NMPC approach can achieve better performances than the manual case and guarantee nearly the same optimality level as the ideal impulsive strategy, presenting however several advantages. In particular, NMPC can allow a significantly higher level of autonomy and offer more flexibility and adaption capability with respect to traditional approaches. In future works, the aim will be to develop NMPC with low-thrust policies, for allowing relevant reductions of fuel consumption.