Model Evaluation of a Transformable CubeSat for Nonholonomic Attitude Reorientation Using a Drop Tower

Yuki Kubo, Tsubasa Ando, Hirona Kawahara, Shu Miyata, Naoya Uchiyama, Kazutoshi Ito, Yoshiki Sugawara · 2025

Read Model Evaluation of a Transformable CubeSat for Nonholonomic Attitude Reorientation Using a Drop Tower by Yuki Kubo, Tsubasa Ando, Hirona Kawahara, Shu Miyata, Naoya Uchiyama, Kazutoshi Ito, Yoshiki Sugawara (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-17173.txt.

Introduction

Yuki Kubo and co-authors design in this 2025 paper a drop tower test for a shape-shifting small satellite. Their transformable CubeSat turns itself in space by flexing joints in sequence, a trick called nonholonomic attitude control, with no thrusters or wheels. A 3U-sized mock-up robot was built to fit the few seconds of free fall a drop tower grants.

The joint sequence was tuned to squeeze the most turning into that blink. Sensors logged spin rates for all four body segments, and the math model matched the tumble well; small fixes after the test sharpened it further. The authors close by mapping what orbit trials will add.

Catalog record

CREATORYuki Kubo, Tsubasa Ando, Hirona Kawahara, Shu Miyata, Naoya Uchiyama, Kazutoshi Ito, Yoshiki Sugawara
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE48 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-17173

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Model Evaluation of a Transformable CubeSat for Nonholonomic Attitude Reorientation Using a Drop Tower Yuki Kubo, Tsubasa Ando, Hirona Kawahara, Shu Miyata, Naoya Uchiyama, Kazutoshi Ito, Yoshiki Sugawara Abstract This paper presents a design for a drop tower test to evaluate a numerical model for a structurally reconfigurable spacecraft with actuatable joints, referred to as a transformable spacecraft. A mock-up robot for a 3U-sized transformable spacecraft is designed to fit within the limited time and space constraints of the microgravity environment available in the drop tower. The robot performs agile reorientation, referred to as nonholonomic attitude control, by actuating joints in a particular manner. To adapt to the very short duration of microgravity in the drop tower test, a successive joint actuation maneuver is optimized to maximize the amount of attitude reorientation within the time constraint. The robot records the angular velocity history of all four bodies, and the data is analyzed to evaluate the accuracy of the numerical model. We confirm that the constructed numerical model sufficiently replicates the robot’s motion and show that the post-experiment model corrections further improve the accuracy of the numerical simulations. Finally, the difference between this drop tower test and the actual orbit demonstration is discussed to show the prospect. [type=editor, auid=000,bioid=1, orcid=0000-0001-7710-9504] [] keywords Drop tower test ,Reconfigurable spacecraft ,Attitude reorientation ,Nonholonomic system ,Particle swarm optimization ,CubeSat † † titlenote: This document is the result of the research project funded by the Grant for a strategic research group from the Advisory Committee for Space Engineering in Japan. † † titlenote: The preprint version of this paper was published online at https://doi.org/10.48550/arXiv.2501.17173 † † credit: Conceptualization of this study, Methodology, Data curation, Project administration, Writing, Investigation † † credit: Data curation, Software, Investigation † † credit: Data curation, Software, Investigation † † credit: Investigation † † credit: Investigation † † credit: Investigation, Methodology, Software † † credit: Supervision † † corresponding: Corresponding author 1 Introduction Novel space exploration technologies have grown rapidly and widely in recent years. Future space robots (including spacecraft in a broad sense) will need to be, for example, more autonomous, more intelligent, more adaptive, more dexterous, and more economical. A transformable spacecraft is one promising architecture, enabling a spacecraft to reconfigure itself with actuatable joints. We can expect the transformable spacecraft to perform multi-objective operations with highly redundant actuators. The Transformer research group, organized by JAXA and universities in Japan, has proposed and studied a concept of such a highly redundant ”Transformer” [ 1 ] . The other group, Tokyo Institute of Technology, launched HIBARI in 2021 [ 2 , 3 ] . HIBARI had only four degrees of freedom actuators, but it carried out important and insightful orbit demonstrations with actuatable mechanisms. One key technology for the transformable spacecraft is ”nonholonomic attitude control,” where the spacecraft can reorient its attitude by successive joint actuation. The study of nonholonomic systems has a long history. Comprehensive surveys can be found in textbooks such as that by Bloch [ 4 ] . As for the mechanics of free-flyers, one of the most widely known nonholonomic problems is the righting reflex of falling cats (or the falling-cat phenomenon). This is a motion in which a cat can always land in an upright attitude by twisting its body in the air, even when dropped in an upside-down attitude. There have been many studies to elucidate the dynamics of this phenomenon [ 5 ] [ 6 ] . Since the 1980s, when numerous space robots were proposed to construct space stations [ 7 ] [ 8 ] , more studies have focused on the dynamics of free-flying robots. Some researchers have proposed actively using the attitude reaction when a manipulator is actuated. Many attitude reorientation methods have been proposed that leverage the nonholonomy of the attitude motion [ 9 ] [ 10 ] [ 11 ] [ 12 ] [ 13 ] [ 14 ] [ 15 ] [ 16 ] [ 17 ] . Physical properties such as mass, position of the center of mass, and moment of inertia determine the direction and amount of the reorientation of the nonholonomic attitude control . Therefore, understanding the fidelity and accuracy of the numerical model is important for the maneuver design. In general, a spacecraft is composed of many components and is not necessarily precisely modeled due to time and cost constraints. Moreover, some components, such as flexible harnesses and liquid propellants, cannot be modeled precisely. Thus, measuring such physical properties through ground tests and operations in orbit is essential. Hardware-based experiments are essential to evaluate the modeling accuracy of the physical properties of robots. There are several examples of experiments to demonstrate the nonholonomic attitude control in hardware. Some researchers attempted to apply the righting reflex of falling cats to free-flying robots or spacecrafts [ 18 ] [ 19 ] [ 20 ] . The most active projects in recent years are carried out by ETH Zurich, named ”SpaceBok” [ 21 ] and ”SpaceHopper” [ 22 ] . Both are legged robots designed to walk and jump in a low-gravity space environment. After a jump on the surface, the robot changes its attitude by nonholonomic attitude control using its legs. To construct the reorientation control law, they adopted deep reinforcement learning in a simulation environment. The simulation parameters, such as masses, moment of inertia, and friction coefficients, cannot be precisely modeled in the simulator, and thus, they used the method of domain randomization [ 23 ] . The policy was trained on many environments in parallel, each having a randomized set of simulation parameters with 20–50% uncertainty. However, such uncertainty may deteriorate the capability of attitude control. In this situation, evaluating and updating such physical parameters using the obtained experimental data can contribute to enhancing attitude control capability. This article describes a method of model evaluation using a drop tower facility. We introduce the experiment’s design and analyze the obtained results. The motivation of the experiment is to estimate the modeling error and the performance of attitude reorientation. Moreover, by extrapolating the result obtained in the drop tower test, we can estimate the on-orbit reorientation performance assuming a certain 3U transformable CubeSat. Section 2 describes the facility of the drop tower, the design of a transformable robot, and a release mechanism. Section 3 first introduces an overview of the nonholonomic attitude control and is followed by physical modeling and maneuver designs using the particle swarm optimization. Section 4 explains setups of the experiment and is followed by results in Section 5 . Finally, Section 6 discusses expected on-orbit performance and several differences between the ground and on-orbit experiment. The preprint version of this article can be found in the arXiv server [ 24 ] . 2 Design of hardware This section describes the experiment facility and the hardware design. 2.1 Facility of the drop tower A drop tower is less costly than an orbit demonstration and a parabolic flight by an airplane. In addition, the test hardware can be refurbished on the ground during relatively long intervals between capsule drops. For these reasons, we adopted the drop tower as a test facility for this experiment. In particular, the drop tower facility COSMOTORRE in Hokkaido, Japan,

Read the complete text