🔭 astrophysics

Multi-sensor fusion for fine-guidance and milliarcsecond-level attitude estimation of balloon-borne telescope

This paper characterizes roll leakage in balloon-borne telescopes using SuperBIT flight data and demonstrates through simulation that multi-star guidance architectures significantly reduce residual image motion compared to single-star systems, motivating their adoption for the planned GigaBIT successor.

Original authors: Philippe Voyer, Maya Amit, Steven J. Benton, Benjamin E. Boyd, Anthony M. Brown, Giulia Cerini, Paul Clark, Matthew Craigie, Christopher J. Damaren, Tim Eifler, Spencer W. Everett, Aurelien A. Fraisse

Published 2026-08-03

📖 3 min read☕ Coffee break read

CC BY 4.0

Original authors: Philippe Voyer, Maya Amit, Steven J. Benton, Benjamin E. Boyd, Anthony M. Brown, Giulia Cerini, Paul Clark, Matthew Craigie, Christopher J. Damaren, Tim Eifler, Spencer W. Everett, Aurelien A. Fraisse, Leo W. H. Fung, Ajay S. Gill, Suren Gourapura, Eric Habjan, John W. Hartley, David Harvey, Bradley Holder, Eric M. Huff, Mathilde Jauzac, William C. Jones, David Lagattuta, Gavin Leroy, Jason S. -Y. Leung, Lun Li, Thuy Vy T. Luu, Richard Massey, Jacqueline E. McCleary, Adyn Miles, Johanna M. Nagy, C. Barth Netterfield, Emaad Paracha, Susan F. Redmond, Jason D. Rhodes, Andrew Robertson, L. Javier Romualdez, Sayan Saha, Jürgen Schmoll, Mohamed M. Shaaban, Ellen Sirks, Sut Ieng Tam, Simon Tartakovsky, Georgios N. Vassilakis, André Z. Vitorelli, Bryce Warren, Alfredo Zenteno

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine trying to take a perfect photograph of a distant galaxy, but your camera is hanging from a giant, wobbly party balloon floating in the stratosphere. The balloon sways like a pendulum, and the whole structure vibrates, making the image blur. To fix this, astronomers use a "Fine Guidance System" (FGS). Think of this system as a super-smart, ultra-fast hand holding the camera lens. Inside the camera, there's a tiny, magical mirror called a Fast Steering Mirror (FSM). If the camera starts to shake, sensors spot the movement, and the FSM tilts instantly to counteract it, keeping the stars sharp.

However, there's a tricky catch. The FSM is like a hand that can only move left-right and up-down. It can't twist or roll. If the balloon starts to spin slightly around the line of sight (like a top wobbling), the FSM can't stop the spin itself. Instead, it tries to "pin" one specific star in the middle of the view to stop it from moving. But because the whole field is actually rotating, fixing that one star makes the stars at the edges of the photo start to smear and stretch. This sneaky problem is called "roll leakage." It's like trying to keep a spinning pizza dough flat by holding down just the center; the edges will still fly off. For telescopes that want to see huge, wide slices of the sky, this smearing is a big deal because it ruins the clarity of the beautiful cosmic pictures they are trying to capture.

This paper, written by a team of scientists and engineers, investigates exactly how bad this "roll leakage" is and how to fix it. They started by looking at real data from a telescope called SuperBIT, which flew on a balloon in 2023. They found that even when the system was doing a great job keeping the main guide star steady, the rest of the image was still wobbling because of that invisible roll. It was like watching a dancer spin: if you only watch their nose, it looks still, but their feet are flying around.

To solve this, the team built a computer simulation to test a new idea: instead of using just one star to tell the mirror how to move, what if we use many stars at once? Imagine trying to balance a spinning plate. If you only watch the center, you might miss the wobble at the edge. But if you watch five different points on the plate simultaneously, you can figure out exactly how it's spinning and how to stabilize the whole thing. The researchers simulated this "multi-star" approach using the design of SuperBIT and a planned, much larger telescope called GigaBIT.

Their results were quite promising. When they used the old "one-star" method, the image smearing was significant. But when they switched to the "multi-star" method in their simulations, the smearing dropped dramatically. For the current SuperBIT telescope, using multiple stars reduced the image motion by about 32%. But for the future, giant GigaBIT telescope, the improvement was massive: a 77.4% reduction in smearing. The paper suggests that for these wide-field balloon telescopes, looking at many stars at once is a much better way to keep the whole picture sharp, rather than just focusing on one. While these findings come from computer models and past flight data rather than a new flight test, they strongly point toward a new way of guiding telescopes to get clearer views of the universe.

Technical Summary: Multi-sensor Fusion for Fine-Guidance and Milliarcsecond-Level Attitude Estimation of Balloon-Borne Telescopes

Problem Statement
Balloon-borne telescopes, such as SuperBIT and the planned GigaBIT, rely on dual-stage pointing architectures to achieve milliarcsecond-level image stability. This architecture combines a coarse three-axis gimbal system with a Fine Guidance System (FGS) utilizing a Fast Steering Mirror (FSM). While the FSM effectively corrects tip and tilt (translational jitter) in two focal-plane axes, it lacks direct control authority over rotation about the boresight (roll). Consequently, residual boresight roll induces field-dependent image motion (field rotation) across the focal plane.

When the FGS locks onto a single guide star, the FSM stabilizes that specific location by interpreting the roll-induced motion as a tip-tilt error. However, this "pins" the focal plane at the guide star, leaving residual motion at other locations on the focal plane. This phenomenon, termed roll leakage, results in systematic, field-dependent image motion that degrades the Point Spread Function (PSF) over the science field, particularly for wide-field instruments like GigaBIT where the distance from the guide star to the science field is larger. The paper identifies that relying on a single guide star and inertial sensors (gyroscopes) is insufficient to fully characterize and correct this roll-induced motion, as gyroscopes fail to capture structural flexure and optical dynamics that contribute to focal-plane motion.

Methodology
The authors employ a three-pronged approach combining flight data analysis, simulation framework development, and comparative architecture testing:

Flight Data Analysis (SuperBIT 2023): The team analyzed data from the 2023 SuperBIT science flight. They utilized a configuration where one Focal-Plane Star Camera (FSC) served as the active guide sensor for the FSM, while a second FSC acted as a passive sensor. By comparing the centroid motion of the passive FSC against independent roll measurements from a Roll Star Camera (RSC) and FSM commands, they quantified the correlation between boresight roll and residual focal-plane motion.
Simulation Framework: A closed-loop simulation was developed incorporating:
- Optical Ray Tracing: Propagating rays through the FSM to the focal plane based on SuperBIT and GigaBIT geometries.
- Disturbance Modeling: Injecting representative low-frequency pendulation and roll disturbances derived from SuperBIT flight Power Spectral Densities (PSD).
- Estimation & Control: Implementing two estimation architectures: (a) Single-star, using only the active guide star, and (b) Multi-star, fusing measurements from multiple asynchronous FSCs using a weighted average approach where weights decay with measurement age. The FSM is controlled via Proportional-Integral (PI) loops.
Comparative Testing: The framework was used to simulate roll-only disturbances for both SuperBIT and a representative GigaBIT geometry. Performance was evaluated by calculating the average Root Mean Square (RMS) image motion across the science field.

Key Contributions

Characterization of Roll Leakage: The study provides empirical evidence from the SuperBIT 2023 flight that residual focal-plane motion is strongly correlated with boresight roll. Analysis of 2,348 science exposures showed that passive FSC motion frequently correlates with RSC roll measurements ( $|\rho| > 0.5$ in most cases), confirming that roll leakage is an inherent limitation of single-guide-star architectures.
Quantification of Amplification: The authors found that the coupling between RSC roll and passive FSC motion exceeds the rigid-body geometric prediction by a factor of approximately 27. This suggests significant amplification due to field-dependent tip-tilt projection or unmodeled structural flexure.
Multi-Star Estimation Framework: The paper introduces a simulation framework that models the coupling between boresight roll, guide-star location, and FSM correction, allowing for the evaluation of multi-star guidance strategies.

Results

SuperBIT Geometry: In simulations using SuperBIT's optical layout, the multi-star estimator reduced the average science-field image motion by 31.8% compared to the single-star architecture. While the single-star approach stabilized the guide star effectively, it resulted in elongated image trails across the rest of the field. The multi-star approach distributed the residual motion more uniformly, effectively shifting the center of rotation toward the center of the science field.
GigaBIT Geometry: When applied to the larger aperture and focal plane of GigaBIT, the improvement was significantly more pronounced. The multi-star estimator reduced average science-field image motion by 77.4%. The authors note that the single-star solution for GigaBIT would result in an average motion of 2.343 arcseconds, whereas the multi-star solution reduced this to 0.529 arcseconds.
Mechanism: The multi-star approach sacrifices perfect stabilization at a single guide-star location to minimize the weighted average motion across the entire science field. By fusing measurements from stars distributed across the field, the estimator can distinguish between translational jitter and rotational roll, allowing the FSM to correct the global motion rather than just the local guide star.

Significance and Claims
The paper claims that roll leakage is a critical, inherent limitation for wide-field balloon-borne telescopes that cannot be fully mitigated by coarse pointing or single-star fine guidance alone. The results motivate the adoption of multi-star fine-guidance architectures for future missions, specifically GigaBIT.

The authors assert that as balloon observatories scale to larger apertures and wider fields of view (like GigaBIT), the penalty for single-guide-star control becomes prohibitive. The multi-star approach offers a scalable solution that leverages the larger collecting area and focal plane of next-generation telescopes to access more guide stars, thereby enabling the milliarcsecond-level stability required for diffraction-limited observations. The paper concludes that combining multi-star fine guidance with improved coarse roll rejection at the gimbal level provides a viable path toward meeting GigaBIT's stringent 0.020 arcsecond image stability requirement.

Technical Summary: Multi-sensor Fusion for Fine-Guidance and Milliarcsecond-Level Attitude Estimation of Balloon-Borne Telescopes

More like this