Sensitivity Analysis for an On-Orbit Robotic Approach Task

Kurzfassung

In application areas such as spacecraft guidance and robotics, optimal control typically requires feasibility with respect to motion constraints, in addition to optimality. Inevitably, the question arises of how to manage uncertainty in related task parameters. Via the Sensitivity Theorem, solutions of a given parametric optimal control problem can be approxi- mated through a fast online sensitivity-based update of a given nominal solution for perturbed task parameter values. Using the presented estimation of the neighborhood of validity of this approximation, a region of task parameter space -- called the task workspace -- can be developed, within which it is possible to provide admissible solutions that are provably robust to task uncertainty. The method is demonstrated in simulation for an on-orbit robotic approach task, which precedes the capture of a target satellite. The non-holonomic kinematically redundant free-floating space robot executes the task under parametric uncertainty of the final six-dimensional pose of the end-effector. This simulation shows that the task workspace is a useful real-world tool for space mission design. The computational efficiency of the solution update within the approximated neighborhood is also demonstrated.