Sliding motions on systems with non-Euclidean state spaces: A differential-geometric perspective
Fernando Castaños · 2025
Read Sliding motions on systems with non-Euclidean state spaces: A differential-geometric perspective by Fernando Castaños (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-2504-19504.txt.
Introduction
Fernando Castaños extends sliding-mode control theory to systems whose states live on curved spaces rather than flat ones. Sliding-mode control is a robust method that forces a system's state to slide along a chosen surface. Many real systems, such as robots with rotating joints, move on smooth curved spaces called manifolds, where flat-space math does not directly apply.
With differential geometry, Castaños rebuilds the basic notions of the theory for this setting. He restates what a solution means for systems that switch behavior, and he gives a geometric account of higher-order sliding modes. He also generalizes the regular form used in design, and writes out first- and second-order controllers that respect the curved structure. Examples on a cylinder, a Möbius bundle, and a sphere show how topology itself can block or shape sliding motion.
Catalog record
| CREATOR | Fernando Castaños |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-19504
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Sliding motions on systems with non-Euclidean state spaces: A differential-geometric perspective Fernando Castaños Abstract This paper extends sliding-mode control theory to nonlinear systems evolving on smooth manifolds. Building on differential geometric methods, we reformulate Filippov’s notion of solutions, characterize well-defined vector fields on quotient spaces, and provide a consistent geometric definition of higher-order sliding modes. We generalize the regular form to non-Euclidean settings and design explicit first- and second-order sliding-mode controllers that respect the manifold structure. Particular attention is given to the role of topological obstructions, which are illustrated through examples on the cylinder, Möbius bundle, SO(3), and 2-sphere. Our results highlight how geometric and topological properties fundamentally influence sliding dynamics and suggest new directions for robust control on nonlinear spaces. 1 Introduction Differential geometry provides the natural mathematical framework for describing and analyzing nonlinear control systems whose state spaces are smooth manifolds rather than Euclidean spaces [ 7 , 1 , 13 ] . Many control problems—such as rigid-body stabilization, spacecraft attitude control, and mobile robotics—involve configuration spaces with rich geometric and topological structures [ 26 , 7 ] . Concepts such as vector fields, flows, tangent bundles, and Lie group actions are indispensable for understanding the behavior of such systems beyond the scope of traditional Euclidean analysis [ 18 ] . In nonlinear control theory, differential geometric methods have profoundly advanced the study of controllability, feedback equivalence, and stabilization [ 15 , 13 , 26 , 24 , 6 ] . They have also revealed fundamental limitations, such as the impossibility of global asymptotic stabilization on certain manifolds due to topological obstructions [ 3 , 14 ] . As engineering applications continue to push the boundaries of performance in geometrically constrained settings, there is an increasing need to integrate geometric insights into controller design. Sliding-mode control is a powerful and robust technique for stabilizing nonlinear systems, particularly in the presence of model uncertainties and external disturbances. Classical formulations of sliding-mode control were developed for systems evolving on Euclidean spaces, where sliding manifolds are typically linear or affine, and discontinuous feedback can be readily implemented [ 30 , 27 ] . Extending sliding-mode control to systems evolving on smooth manifolds introduces new challenges: switching surfaces must be modeled as embedded submanifolds, vector fields must respect the manifold structure, and topological constraints can fundamentally alter the nature of sliding dynamics. With few exceptions—most notably the early and pioneering work of Sira-Ramírez [ 28 ] —differential geometric methods have not been systematically developed within the context of sliding-mode control. In particular, the geometric and topological aspects of switching manifolds, sliding vector fields, and higher-order sliding motions on manifolds remain largely unexplored. 1.1 Contributions In this paper, we extend sliding-mode control theory to systems whose state spaces are smooth manifolds. We characterize well-defined maps and vector fields on quotient manifolds—a result with potential applications beyond the sliding-mode control community. We reformulate Filippov’s notion of solutions in a geometric setting, provide a consistent definition of higher-order sliding modes, and point out limitations in some existing formulations [ 32 , 31 ] . We also generalize the concept of regular form to non-Euclidean spaces and construct explicit first- and second-order sliding-mode controllers that respect the geometry of the underlying manifold. Special attention is given to topological obstructions, illustrated through examples on the cylinder, Möbius bundle, \mathrm{SO}(3) , and 2-sphere. Our results underscore the importance of geometric and topological considerations in robust control design and suggest new directions for sliding-mode control on general nonlinear spaces. 1.2 Paper Structure Section 2 recalls the necessary background for discussing dynamical systems on smooth manifolds and characterizes well-defined maps and vector fields on quotient spaces. Geometric definitions of sliding motions are introduced in Section 3 . In Section 4 , we generalize the regular form and provide design examples of second-order sliding-mode controllers on non-Euclidean state spaces. Finally, concluding remarks are presented in Section 5 . 2 Global representations for non-Euclidean state-spaces In the differential-geometric approach to dynamical systems, it is assumed that the state-space is a smooth n -dimensional manifold X and that the system’s motions are governed by a vector field taking values on its tangent bundle TX . More precisely, the system is described by differential equations of the form \dot{x}=f(x,t)\;, (1) where x\in X is the state, t\in\mathbb{R} is time, and f(x,t) belongs to the tangent space T_{x}X of X at x . Since manifolds are formally defined through local coordinate charts, the geometric objects on such manifolds (e.g., maps, vector fields, or differential forms) are usually specified by patching local coordinate-based definitions. A usual local model for the system is thus \dot{\xi}=\varphi(\xi,t)\;, (2) where \xi\in W\subset\mathbb{R}^{n} is the state coordinate, that is, \xi=\psi(x) with \psi:U\to W a homeomorphism and U\subset X a local coordinate domain. Here, \varphi:W\times\mathbb{R}\to\mathbb{R}^{n} is a local coordinate-based definition of the time-dependent vector field f . Unfortunately, using multiple coordinate charts in control applications is impractical, so the designer sometimes settles for a local controller. There are cases, however, in which a globally defined control law is essential, and a coordinate-free representation of X becomes mandatory. Throughout, x will denote a point of the abstract manifold X , while \xi will denote its coordinate representation in a local chart. This section explores two extrinsic ways to construct a coordinate-free representation of a state space. In order to define and design sliding motions, we will be required to construct subsets, maps, and vector fields on such state spaces. 2.1 Embedded manifolds Allow us to begin by recalling some definitions. If X and \tilde{X} are smooth manifolds and \Psi:X\to\tilde{X} is a smooth map, for each x\in X we define a map \Psi_{*}:T_{x}X\to T_{\Psi(x)}\tilde{X} , called the pushforward associated with \Psi , by (\Psi_{*}v)g=v(g\circ\Psi) , for any tangent vector v\in T_{x}X and any smooth function g:\tilde{X}\to\mathbb{R} . A smooth map \Psi:X\to\tilde{X} is called a submersion if \Psi_{*} is surjective at each point (or equivalently, if \rank\Psi=\dim\tilde{X} ). It is called an immersion if \Psi_{*} is injective at each point (equivalently, if \rank\Psi=\dim X ). A smooth embedding is a smooth map \Psi:X\to\tilde{X} that is both an immersion and a topological embedding. That is, it maps a manifold X smoothly into another manifold \tilde{X} such that the image \Psi(X) inherits the subspace topology from \tilde{X} , and the map is a homeomorphism onto its image. An embedded submanifold of a manifold \tilde{X} is a subset X\subset\tilde{X} that forms a manifold under the subspace topology. It is equipped with a smooth structure so that the natural inclusion map X\hookrightarrow\tilde{X} qualifies as a smooth embedding. See [ 18 ] for details. One way to obviate the need for multiple charts is to embed X in a higher-dimensional real space \mathbb{R}^{p} . While this is always possible due to the Whitney Embedding Theorem [ 18 , Thm. 6.15] , we note that the proof of the Theorem is not constructive. Fortunately, explicit embeddings are known for a variety of manifolds relevant to applications (the sphere, the torus, SO(3), real projective space, etc). Example 1 (The Group of Rotations \mathrm{SO}(3) ) . In the attitude control problem for rigid bodies, the configuration manifold is the group \mathrm{SO}(3) of all rotations about the origin in \mathbb{R}^{3} . An effective global