Data-Enabled Predictive Control for Flexible Spacecraft

Huanqing Wang, Kaixiang Zhang, Amin Vahidi-Moghaddam, Haowei An, Nan Li, Daning Huang, Zhaojian Li · 2025

Read Data-Enabled Predictive Control for Flexible Spacecraft by Huanqing Wang, Kaixiang Zhang, Amin Vahidi-Moghaddam, Haowei An, Nan Li, Daning Huang, Zhaojian Li (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-2502-09531.txt.

Introduction

Huanqing Wang and co-authors steer flexible spacecraft using flight data instead of a hand-built model. Their 2025 paper addresses the bendy panels and appendages that light spacecraft carry to save weight. Such parts flex in hard-to-model ways, which makes precise control tricky.

The method, called data-enabled predictive control, builds its guide directly from recorded past motion. A classic result in control theory lets those records stand in for a model. Dimension reduction keeps the math light enough to run. In simulations the data-driven method beat a Lyapunov-based controller, the authors report. They present the work as a first use of this framework for boundary control of flexible craft.

Catalog record

CREATORHuanqing Wang, Kaixiang Zhang, Amin Vahidi-Moghaddam, Haowei An, Nan Li, Daning Huang, Zhaojian Li
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE55 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-09531

Browse the full library · About the Tower of Babel

Full text

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

Data-Enabled Predictive Control for Flexible Spacecraft Huanqing Wang, Kaixiang Zhang, Amin Vahidi-Moghaddam, Haowei An, Nan Li, Daning Huang, Zhaojian Li Abstract Spacecraft are vital to space exploration and are often equipped with lightweight, flexible appendages to meet strict weight constraints. These appendages pose significant challenges for modeling and control due to their inherent nonlinearity. Data-driven control methods have gained traction to address such challenges. This paper introduces, to the best of the authors’ knowledge, the first application of the data-enabled predictive control (DeePC) framework to boundary control for flexible spacecraft. Leveraging the fundamental lemma, DeePC constructs a non-parametric model by utilizing recorded past trajectories, eliminating the need for explicit model development. The developed method also incorporates dimension reduction techniques to enhance computational efficiency. Through comprehensive numerical simulations, this study compares the proposed method with Lyapunov-based control, demonstrating superior performance and offering a thorough evaluation of data-driven control for flexible spacecraft. Index Terms —Flexible spacecraft, data-driven control, predictive control. I Introduction Flexible spacecraft have attracted considerable interest due to their wide range of applications, including space missions, communication, and remote sensing [ 1 ] . The high cost of launches has driven the demand for lighter structures and compact servomechanisms. To minimize weight, spacecraft components such as solar panels and antennas are often designed as lightweight and flexible structures. Spacecraft such as Voyager 1 feature long, extended appendages that position instruments, such as magnetometers, farther from the spacecraft to reduce electromagnetic interference. These long structures, often equipped with sensors, are oriented at specific angles to ensure accurate measurements or signal transmission. However, this flexibility induces vibrations that can impair functionality and performance. Prolonged undesirable vibrations can reduce the spacecraft’s lifespan and even lead to failure. For complex systems like flexible spacecraft, the dynamics are often represented by partial differential equations (PDEs) coupled with ordinary differential equations (ODEs) [ 2 , 3 ] . Various control methods have been investigated for such systems, including proportional-derivative (PD) control [ 4 ] , linear-quadratic regulator [ 5 ] , adaptive control [ 2 ] , variable structure control [ 3 ] , and robust control [ 6 ] . In particular, adaptive control is applied in [ 7 ] to a flexible moving beam, and [ 8 ] utilizes model reference adaptive control to manage a flexible appendage with a rigid central body. Due to the complexity of controlling systems governed by PDEs, a common approach is to discretize the PDE system, reducing it to finite-dimensional ODEs. Then, controllers are developed based on the resulting ODEs; see e.g., [ 3 ] . The assumed mode method is popular for this purpose, and extensive control methods have been applied to the ODEs generated from this approach [ 9 , 10 ] . However, this model reduction can lead to spillover instability [ 11 , 12 ] , which occurs when controller is not designed to compensate for higher or unmodeled modes. In recent years, research has shifted towards direct PDE boundary control to address these issues. Boundary control, in which actuation and sensing occur exclusively through the boundary conditions, applies control effort at the boundary to guide the appendage to its desired position while simultaneously suppressing unwanted vibrations. The authors of [ 13 ] develop a single control input at the center of the body for a flexible solar panel modeled as two Euler-Bernoulli beams. Meng [ 14 ] derives a Lyapunov barrier function that guarantees convergence while maintaining constraints, and Rad [ 15 ] further advances this work by achieving angle and vibration control without relying on system damping. Additionally, Ma [ 16 ] proposes a similar PDE control scheme using the Lyapunov direct method. The commonality of these approaches is Lyapunov-based control, which derives control laws from the system model by selecting a Lyapunov candidate function that guarantees stability and convergence. The Lyapunov candidate function typically considers the total system energy, tracking performance, and exponential stability. One drawback of Lyapunov-based control is the intensive derivation process required to formulate the control law, which relies on an accurate pre-derived model. To overcome these limitations, data-driven control has gained increasing attention for handling complex and nonlinear systems. Unlike traditional methods, data-driven approaches do not require detailed knowledge of the underlying physics. One promising technique in this field is D ata- e nabl e d p redictive c ontrol (DeePC), which leverages the fundamental lemma [ 17 , 18 ] , asserting that for a controllable linear time-invariant (LTI) system, any trajectory of the system can be represented as a linear combination of sequences obtained from a persistently exciting input-output data. DeePC develops non-parametric models from data without the need for explicit system models. This technique has been applied successfully to various applications, including power systems [ 19 ] , energy management [ 20 ] , quadcopters [ 21 ] , soft robots [ 22 ] , and vehicle platooning [ 23 ] . The ability to generate control inputs directly from trajectory data makes DeePC particularly suitable for systems with complex, hard-to-model dynamics, such as flexible spacecraft. This paper makes the following three primary contributions. First, it presents the first application of DeePC to the control of flexible spacecraft that circumvents the need for complicated modeling and controller derivation. Second, it demonstrates the capability of DeePC to track angles and stabilize an infinite number of vibration modes without requiring parametric models or knowledge of system parameters. Third, the proposed method is compared with the well-received Lyapunov-based control method under nominal conditions, uncertainty conditions, and process noise conditions, using a finite element (FE) model. The remainder of this paper is structured as follows. Section II provides an overview of the modeling of flexible spacecraft. Section III introduces the background of DeePC and its application to the specific problem. Section IV describes the benchmark approaches and compares them with DeePC in numerical simulation. Finally, Section V concludes this study. II System Description In this section, we present the system description and mathematical modeling of the flexible spacecraft. Since DeePC is a data-driven method, the model will be solely used to generate output data from input data, and the underlying dynamics will not be utilized by DeePC. The flexible spacecraft of interest is a satellite with a long, flexible antenna. The spacecraft can be modeled as a central hub mass with a flexible cantilever beam, as illustrated in Fig. 1 . The flexible beam is modeled as an Euler-Bernoulli beam. The control objective is to guide the flexible beam to a desired angle \theta_{d} while suppressing unwanted vibrations. Due to the material’s flexibility, the beam deforms during motion, making it challenging to accurately track the angle. In the absence of gravity and friction in space, the flexible beam may vibrate indefinitely. The coordinate frames are defined as follows: XOY is the inertial coordinate frame, and X_{b}O_{b}Y_{b} is the body-fixed coordinate frame. The pitch angle of rotation of the hub is denoted by \theta(t) , and the elastic displacement of the beam at position x with respect to the body frame is represented by y(x,t) . The position of a point on the appendage in the inertial frame is given by z(x,t)=x\theta(t)+y(x,t) (Note: the radius of the hub, r

Read the complete text