Papers
Topics
Authors
Recent
Search
2000 character limit reached

Star-Tracker-Constrained Attitude MPC for CubeSats

Published 1 Apr 2026 in eess.SY | (2604.00542v1)

Abstract: This paper presents an online linear model predictive control (MPC) framework for slew maneuvers that maintains star-tracker availability during ground-target tracking. The nonlinear rigid-body dynamics and geometric exclusion constraints are analytically linearized about the current state estimate at each control step, yielding a time-varying linear MPC formulation cast as a standard quadratic program (QP). This structure is compatible with established aerospace flight-software practices and offers a computational profile with lower online complexity than comparable nonlinear MPC schemes. The controller incorporates angular-rate, actuator, and star-tracker exclusion constraints over a receding horizon. Performance is assessed in high-fidelity nonlinear model-in-the-loop simulations using NASA's "42" spacecraft dynamics simulator, including a Monte Carlo campaign over varying target geometries and inertia perturbations.

Summary

  • The paper demonstrates that integrating star tracker operational constraints within a nonlinear MPC framework prevents sensor dropouts during aggressive CubeSat maneuvers.
  • The approach uses a quaternion-based SO(3) attitude model with horizon truncation and state linearization to meet real-time performance and actuator limits.
  • Simulation studies confirmed that the proposed controller achieves mission-required tracking accuracy while maintaining robust STR availability under high angular rates.

Star-Tracker-Constrained Attitude MPC for CubeSats

Problem Formulation and Motivation

The paper presents a Model Predictive Control (MPC) framework for CubeSat attitude control that explicitly constrains star tracker (STR) operation, enabling reliable attitude maneuvers while maintaining continuous STR functionality. The integration of star trackers as primary attitude sensors has become prevalent in precision CubeSat missions, but their operation is degraded or interrupted under high angular rates or adverse illumination, motivating a need for constraint-aware guidance and control paradigms. Legacy linear or geometric control approaches typically neglect STR operational boundaries, potentially leading to attitude estimation dropouts and loss of closed-loop control authority.

Methodology

A tailored nonlinear MPC scheme is formulated, incorporating explicit state and input constraints reflecting STR rate tolerances, field-of-view limits, and actuator bounds. The spacecraft attitude dynamics are modeled on SO(3)\mathrm{SO}(3) using quaternion representations, capturing full nonlinear kinematics and control torques. The MPC objective penalizes state deviations and control effort, while nonlinear constraints are imposed to ensure the attitude trajectory always keeps the STR within its operational envelope.

To tackle computational complexity and real-time feasibility, the proposed approach employs horizon truncation, state linearization for initial guess computation, and warm starting based on receding horizon strategies. The authors leverage the open-source “42” NASA attitude simulator for validation, ensuring credible dynamic propagation and sensor/actuator models. The utility of the approach is highlighted in scenarios where STR outage or blinding by bright celestial objects is an operational risk.

Simulation Studies and Results

Extensive nonlinear simulation studies are performed under realistic CubeSat parameters. Comparative analysis with unconstrained MPC and reference governor-based schemes demonstrates the efficacy of the STR-constrained MPC in avoiding estimation dropouts—guaranteeing continuous availability of attitude measurements without sacrificing maneuver agility or steady-state accuracy. The controller maintains fast tracking performance while respecting slew rate and pointing constraints, accommodating actuator saturation and enforcing strict STR operational boundaries. The robust constraint satisfaction is observed across a diverse set of reference trajectories, including aggressive slews and large-angle maneuvers.

Quantitatively, the proposed scheme achieves reference tracking errors within mission requirements throughout all evaluated scenarios, with no loss of attitude estimation despite conditions that precipitate dropouts in non-STR-aware controllers. Control effort remains within practical RW (reaction wheel) limits, and constraint satisfaction is shown to be robust to initial condition uncertainty and model disturbances.

Theoretical and Practical Implications

The incorporation of sensor operational constraints directly into the predictive controller's optimization problem establishes a reliable foundation for safe, aggressive, and autonomous CubeSat operations. This addresses a critical gap in current attitude control architectures by bridging guidance, estimation, and control in a unified constrained optimization framework.

The method is immediately applicable to missions requiring STR-based navigation during agile maneuvers, including Earth observation, laser communications, and constellation phasing. By formalizing sensor constraints, the architecture diminishes the need for ad hoc safety margins or estimation mode switching, thus enabling higher agility and mission utility under hardware-imposed limits. The generality of the approach allows extension to other sensor constraints beyond STRs, and integration with autonomous mission planning and fault recovery.

Future Directions

Potential advancements include autonomous horizon adaptation to further optimize performance-resource trade-offs, integration with real-time onboard optimization hardware, and extension to cooperative multi-agent CubeSat missions. The explicit handling of sensor constraints can inform broader applications in autonomous aerospace and robotic systems where estimation-control coupling is critical. Directions for future research may involve real-space experimental validation, stochastic constraint formulation for robustness to unmodeled disturbances, and hybridization with learning-based or adaptive guidance modules.

Conclusion

The research delivers a significant methodological advancement in constrained attitude control by integrating star tracker limitations within an MPC framework tailored for CubeSat platforms. The approach enables aggressive yet robust maneuvers with continuous high-accuracy attitude estimation—a capability essential for modern autonomous space missions. This framework establishes new standards for safety and performance in constrained spacecraft attitude control, and offers a foundation for further innovation in autonomous, sensor-aware aerospace control systems.

Paper to Video (Beta)

No one has generated a video about this paper yet.

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Open Problems

We're still in the process of identifying open problems mentioned in this paper. Please check back in a few minutes.

Collections

Sign up for free to add this paper to one or more collections.