The Case for Parameter-Aware Control of Assistive Free-Flyers

Abstract

View Video Presentation: https://doi.org/10.2514/6.2021-2018.vid Manipulator-equipped microgravity robots (free-flyers) have been proposed for on-orbit activities including assisting astronauts with daily tasks and serving as mobile workers in microgravity assembly scenarios. An emerging class of free-flyer, assistive free-flyers (AFFs), will engage in complex manipulation scenarios such as grasping and moving new objects. This introduces a notable source of uncertainty in the form of inertial parametric uncertainty stemming from grappling unknown or poorly-characterized objects. This paper explicitly demonstrates the value---and perhaps even necessity---of control that can account for parametric uncertainty in these AFF manipulation scenarios, which involve more agile motion and larger end effector mass fractions than more traditional on-orbit free-flyer control situations. In this case study, a six degree of freedom base AFF with a two-link manipulator grapples an unknown point mass. The free-flying manipulator dynamics are detailed and a complex feasible reference trajectory, representative of those potentially seen by future AFFs, is provided. Two non-adaptive controllers, proportional-derivative (PD) and feedback linearization, are compared with sliding mode adaptive control for varying grappled mass fraction. A Lyapunov-based convergence analysis of the adaptive controller is also explicitly provided. Drawing from the prior free-flying literature, the value of parameter-aware control design for emerging AFF control scenarios is demonstrated.