Architecting Autonomy for Safe Microgravity Free-Flyer Inspection

Abstract

Small free-flying spacecraft can provide vital extravehicular activity (EVA) services like inspection and repair for future orbital outposts like the Lunar Gateway. Operating adjacent to delicate space station and microgravity targets, these spacecraft require formalization to describe the autonomy that a free-flyer inspection mission must provide. This work explores the transformation of general mission requirements for this class of free-flyer into a set of concrete decisions for the planning and control autonomy architectures that will power such missions. Flowing down from operator commands for inspection of important regions and mission time-criticality, a motion planning problem emerges that provides the basis for developing autonomy solutions. Unique constraints are considered such as velocity limitations, pointing, and keep-in/keep-out zones, with mission fallback techniques for providing hierarchical safety guarantees under model uncertainties and failure. Planning considerations such as cost function design and path vs. trajectory control are discussed. The typical inputs and outputs of the planning and control autonomy stack of such a mission are also provided. Notional system requirements such as solve times and propellant use are documented to inform planning and control design. The entire proposed autonomy framework for free-flyer inspection is realized in the SmallSatSim simulation environment, providing a reference example of free-flyer inspection autonomy. The proposed autonomy architecture serves as a blueprint for future implementations of small satellite autonomous inspection in proximity to mission-critical hardware, going beyond the existing literature in terms of both (1) providing realistic system requirements for an autonomous inspection mission and (2) translating these requirements into autonomy design decisions for inspection planning and control.

Figures (6)

• Figure 1: Free-flyer inspection can provide agile, efficient operations in proximity to orbital outposts but will likely require autonomy for maximum safety and optimality. Here, an autonomous inspection framework is depicted in the SmallSatSim simulation environment; red circles indicate waypoints interpolated from manually-specified inspection points, $\boldsymbol{x}_{ins}$ within a purple safety corridor of radius $r_{ins}$.
• Figure 2: $\mathcal{X}_\texttt{out}$ is efficiently constructed by taking convex hulls of major station geometry.
• Figure 3: The layout of the proposed autonomy framework for free-flyer inspection, consisting of inspection points that specify a desired free-flyer pose $\boldsymbol{x}_{\texttt{ins},i}$, which define keep-in corridors $\mathcal{X}_\texttt{in}$ and keep-out zones $\mathcal{X}_\texttt{out}$. The instantaneous radius $r_{ins}$ is adjusted based on the free-flyer radius $r$ and desired safe geometry clearance.
• Figure 4: The proposed autonomy framework for free-flyer inspection planning and control. Note that additional layers of autonomy, like automated trajectory generation and online adaptation, are noted as optional modules. Localization and perception are assumed to be provided by an onboard or offboard global localization system.
• Figure 5: Flyby trajectory tracking in the simulation environment of Figure \ref{['fig:main']}; the red circled portion at right is shown. $\boldsymbol{x}_{ins}$ are user-specified to provide a simple, intuitive inspection interface that can be widely integrated into inspection planning approaches. Note excellent tracking performance even with a single failure (green), and poor performance with severe model degradation (red). Note that linger mode has similar performance to flyby mode and is not shown.