Session
Flash Talk Session 1
Location
Salt Palace Convention Center, Salt Lake City, UT
Abstract
In typical low Earth orbit (LEO) operations, the inability to remove excess angular momentum often leads to the decommissioning of a satellite. This paper presents the on-orbit recovery of GRUS-1E, a hundred kg class Earth observation satellite in Axelspace Corporation’s GRUS-1 constellation. The recovery followed the permanent loss of magnetic torquer commanding capability, which in turn disabled the satellite’s nominal reaction wheel momentum unloading. In addition, the satellite experienced persistent disturbance torques that significantly exceeded the combined contributions predicted by standard models, including gravity gradient, solar radiation pressure, aerodynamic drag, and the estimated residual magnetic dipole moment. The magnitude of these unexplained torques suggested that the residual magnetic dipole moment, or its interaction with Earth’s magnetic field, had changed from its previously characterized value, while the possibility of an additional spacecraft specific disturbance source could not be excluded. These conditions rendered conventional recovery methods, such as thruster-based unloading or closed-loop momentum management, either infeasible or unsustainable. Consequently, the reaction wheels were the only actuators immediately available for recovery.
To address this, we developed an automated ground-to-orbit workflow for attitude-based angular momentum reduction, employing a data-driven approach for disturbance characterization and open-loop attitude steering. The approach centers on two core components: using flight telemetry to characterize the external disturbance, and applying geometric optimal control to generate attitude trajectories that exploit it. Flight telemetry was analyzed using Singular Spectrum Analysis (SSA) and Sparse Identification of Nonlinear Dynamics (SINDy) to rediscover the known deterministic satellite dynamics directly from data. By isolating the residual between identified and theoretical dynamics, we pinpointed the disturbance as an interaction with Earth’s magnetic field, well approximated by a residual magnetic dipole model. This approach facilitated the derivation of an analytical expression for the disturbance torque, allowing forward modeling of previously uncharacterized environmental effects. By exploiting this disturbance, we applied geometric optimal control to generate open-loop target attitude trajectories that enabled passive momentum reduction. The resulting time-tagged attitude command sequences were uploaded and executed without requiring any modifications to the onboard flight software.
On-orbit execution successfully reduced GRUS-1E’s total angular momentum through a fully ground-based control sequence, restoring sufficient control authority to maneuver the satellite for image acquisition and ultimately resume commercial operations as part of the GRUS-1 constellation. These results demonstrate a practical recovery strategy for small satellites, and provide an operational template for leveraging advanced data analysis and control algorithms to extend satellite mission life.
Document Type
Event
Included in
Recovery of GRUS-1E From Magnetic Torquer Loss Using Data-Driven Open-Loop Attitude Steering
Salt Palace Convention Center, Salt Lake City, UT
In typical low Earth orbit (LEO) operations, the inability to remove excess angular momentum often leads to the decommissioning of a satellite. This paper presents the on-orbit recovery of GRUS-1E, a hundred kg class Earth observation satellite in Axelspace Corporation’s GRUS-1 constellation. The recovery followed the permanent loss of magnetic torquer commanding capability, which in turn disabled the satellite’s nominal reaction wheel momentum unloading. In addition, the satellite experienced persistent disturbance torques that significantly exceeded the combined contributions predicted by standard models, including gravity gradient, solar radiation pressure, aerodynamic drag, and the estimated residual magnetic dipole moment. The magnitude of these unexplained torques suggested that the residual magnetic dipole moment, or its interaction with Earth’s magnetic field, had changed from its previously characterized value, while the possibility of an additional spacecraft specific disturbance source could not be excluded. These conditions rendered conventional recovery methods, such as thruster-based unloading or closed-loop momentum management, either infeasible or unsustainable. Consequently, the reaction wheels were the only actuators immediately available for recovery.
To address this, we developed an automated ground-to-orbit workflow for attitude-based angular momentum reduction, employing a data-driven approach for disturbance characterization and open-loop attitude steering. The approach centers on two core components: using flight telemetry to characterize the external disturbance, and applying geometric optimal control to generate attitude trajectories that exploit it. Flight telemetry was analyzed using Singular Spectrum Analysis (SSA) and Sparse Identification of Nonlinear Dynamics (SINDy) to rediscover the known deterministic satellite dynamics directly from data. By isolating the residual between identified and theoretical dynamics, we pinpointed the disturbance as an interaction with Earth’s magnetic field, well approximated by a residual magnetic dipole model. This approach facilitated the derivation of an analytical expression for the disturbance torque, allowing forward modeling of previously uncharacterized environmental effects. By exploiting this disturbance, we applied geometric optimal control to generate open-loop target attitude trajectories that enabled passive momentum reduction. The resulting time-tagged attitude command sequences were uploaded and executed without requiring any modifications to the onboard flight software.
On-orbit execution successfully reduced GRUS-1E’s total angular momentum through a fully ground-based control sequence, restoring sufficient control authority to maneuver the satellite for image acquisition and ultimately resume commercial operations as part of the GRUS-1 constellation. These results demonstrate a practical recovery strategy for small satellites, and provide an operational template for leveraging advanced data analysis and control algorithms to extend satellite mission life.
