Safe On-Orbit Dislodging of Deployable Structures via Robust Adaptive MPC

TL;DR

The paper tackles safe on-orbit dislodging of a jammed solar panel under time-varying client parameters and strict safety constraints. It introduces a robust adaptive MPC (RAMPC) that combines online set-membership parameter estimation with tube-based constraints and a dual-mode cost to balance exploration (learning parameters) and exploitation (achieving precise control). Key contributions include a linear parametric representation of the hinge dynamics, a set-based estimator with fixed-structure online tightening, robust tube and terminal sets, and a convex reformulation enabling real-time MPC; extensive simulations and hardware experiments show improved parameter convergence, tracking accuracy, and safety versus baselines. The approach offers a principled framework for safe, data-efficient on-orbit servicing with time-varying dynamics, potentially enhancing the reliability of autonomous space operations.

Abstract

This paper proposes a novel robust adaptive model predictive controller for on-orbit dislodging. We study orbit dislodging where a servicing spacecraft uses a robotic arm to free a jammed and unactuated solar panel mounted on a hybrid hinge that acts as a time-varying client on a space station. Our method couples online set-membership identification with a robust adaptive MPC to enforce safety under bounded disturbances. The controller explicitly balances exploration to excite the system and shrink uncertainty and exploitation to improve control performance through a dual-mode cost. The feasibility of the developed robust adaptive MPC method is also examined through dislodging simulations and hardware experiments in freefall and terrestrial laboratory environments, respectively. In addition, the advantages of our method are shown through comparison experiments with several state-of-the-art control schemes for both accuracy of parameter estimation and control performance.

Figures (14)

• Figure 1: A servicer docked with a space station dislodges an unactuated solar panel with hinge via its robot arm in space. Supplemental Video: http://tiny.cc/tcst2025unm
• Figure 2: Diagram of the dislodging process that one servicer docks with a space station using its docking system and then dislodges a jamming solar panel that is unactuated via its manipulator in space.
• Figure 3: The block diagram of our robust adaptive MPC algorithm.
• Figure 4: Dislodging mission using a servicer which contains simulated on MuJoCo Platform in a freefall environment.
• Figure 5: Exploitation results for the evolution in both states of the client in which $\theta_k$ in (a) and $\dot \theta_k$ in (b) and control input in which $\mathbf{f}_k$ in (c) and $\boldsymbol{\tau}_k$ in (d) compared to PID, adaptive control and robust adaptive MPC for 40 seconds.

Theorems & Definitions (1)

• proof