Priority-Driven Task Allocation Framework for Robotic On-Orbit Servicing

On-Orbit Servicing (OOS) missions performed by orbital space robots require an optimal and safe system behavior in order to achieve a full stack of multiple tasks. However, OOS tasks may conflict with each other at some particular operational points, especially in case of uncooperative targets with prior unknown motion states and geometry. To address these challenges, we propose a novel task-priority framework for optimal (local) allocation and prioritization of tasks according to user-defined metrics by exploiting the system redundancy within an online algorithm. Our approach uses equality tasks to track state trajectories for the end-effector and optionally also for the servicer-base, and set-based tasks to ensure safety and system constraints such as collision avoidance, the field of view, and manipulator workspace limits. We prove the asymptotic convergence of the equality task errors together with the satisfaction of all set-based tasks. Software-in-the-loop results with an end-to-end simulation test facility show the reliability of the proposed approach aiming to enable more efficient and reliable on-orbit servicing missions.