Modeling and Navigation Filtering for Spacecraft Rendezvous with Flexible Appendages and Sloshing

With applications ranging from autonomous in-orbit servicing, assembly, and manufacturing to orbital debris removal, the complexity of rendezvous operations with space manipulator systems presents significant challenges in enhancing performance while ensuring safety. To enable more agile operations, a key aspect is dealing with flexibility, specifically the deflection of mechanical structures under loads and their subsequent coupling with the Attitude & Orbit Control Subsystem (AOCS). During very close proximity operations, because of the coupling between attitude and translation, applied forces and torques can significantly excite structural flexible modes, which may cause disturbances to the robotic arm’s end-effector and potentially lead to collisions. However, guidance and navigation models typically used in the AOCS are often simplified by using a rigid-body assumption, which does not adequately capture the dynamics of spacecraft with large flexible components (e.g. solar arrays, booms, robotic arms). This assumption becomes more and more obsolete with the development of space missions of the new generation involving very large structures (e.g., space power generators), agile spacecrafts for exploration, or rendezvous operations. Therefore, on the navigation side, assuming a rigid model could lead to unacceptable errors (e.g., tens to hundreds of pixels) between the projection of key points identified by vision algorithms and the actual relative position and orientation between the two spacecraft. Understanding the impact of flexibility on spacecraft navigation in close proximity can thus be key to advance agile in-orbit servicing capabilities. In this context, this paper aims at addressing a significant gap in the literature by taking into account uncertainties and varying parameters in a unique analytical linear model, which can be evaluated at each desired step of the navigation filter. A comprehensive modeling of complex multibody structures allows for early prediction of worst-case scenarios, thereby pushing the system to its performance limits and reducing design margins.