Multiple model trajectory generation for uncertain target spin direction

Abstract

Multiple concepts have emerged that require the docking of two or more spacecrafts, including autonomous robotic servicing and assembly of orbiting assets, and the active removal of space debris. Successful dockings require that the Target's spin rate and direction be determined so that approach trajectories may be computed. Remote sensing from ground stations or distant satellite observations can provide important data for the Chaser satellite, such as the spin rate or amount of tumbling by the Target. Optical light curves are a basic method of determining the period of a Target's rotation. To avoid recomputing the approach trajectory based on updated spin rate or direction data and reduce the computational complexity required to generate a fuel efficient approach trajectory, a means of generating trajectories robust to the spin direction must be developed. This paper presents the multiple model approach, in which training models representing potential spin rates and directions are weighted by their likelihood of representing the true behavior of the Target, for synchronous trajectories. While not fuel or time optimal, the synchronous multiple model trajectory keeps the docking location in the field of view of the Chaser's rendezvous sensors throughout the trajectory while trading performance from the nominal model for robustness across a range of models. Further, the Targets in this analysis are taken to be rigid, uncooperative, and passively tumbling. The paper presents the application of the multiple model approach to docking trajectory generation. In doing so, it compares both the optimal approach trajectory and robust, multiple model trajectory for uncertainty in the spin direction. This paper also assesses several tuning parameters that engineers may use to trade the computational complexity, fuel cost, and robustness of the resulting multiple model trajectory for use in future missions. The processes discussed in the paper therefore can be applied to additional future mission scenarios beyond the case study of a single Chaser approaching a single Target.