Rectified Control Barrier Functions for High-Order Safety Constraints
Pio Ong, Max H. Cohen, Tamas G. Molnar, Aaron D. Ames · 2024
Read Rectified Control Barrier Functions for High-Order Safety Constraints by Pio Ong, Max H. Cohen, Tamas G. Molnar, Aaron D. Ames (2024) 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-2412-03708.txt.
Introduction
Pio Ong, Max Cohen, Tamas Molnar, and Aaron Ames present a safety method in 2024 for control systems. Barrier functions act as fences: a controller that respects them keeps a machine inside safe bounds. Many real constraints sit several steps away from the controls, which makes fence-building hard.
Their rectified barrier functions add an activation term to the classic build, easing the construction. The paper compares the approach with a stepwise alternative and works through three cases of rising difficulty. Plain examples plus an aircraft control problem show the method in action. The result is a cleaner route to provable safety for complex machines.
Catalog record
| CREATOR | Pio Ong, Max H. Cohen, Tamas G. Molnar, Aaron D. Ames |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 38 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-03708
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Rectified Control Barrier Functions for High-Order Safety Constraints Pio Ong, Max H. Cohen, Tamas G. Molnar, Aaron D. Ames Abstract This paper presents a novel approach for synthesizing control barrier functions (CBFs) from high relative degree safety constraints: Rectified CBFs (ReCBFs) . We begin by discussing the limitations of existing High-Order CBF approaches and how these can be overcome by incorporating an activation function into the CBF construction. We then provide a comparative analysis of our approach with related methods, such as CBF backstepping. Our results are presented first for safety constraints with relative degree two, then for mixed-input relative degree constraints, and finally for higher relative degrees. The theoretical developments are illustrated through simple running examples and an aircraft control problem. I Introduction Control barrier functions (CBFs) [ 1 , 2 ] are a useful tool for safety-critical control systems, providing a way to synthesize controllers enforcing state constraints. One of the main advantages of CBFs is the ease of control synthesis using methods such as quadratic programming [ 1 ] , safety filters [ 3 ] , or closed-form solutions [ 4 , 5 ] . Moreover, there exist numerous extensions of CBFs that address properties beyond safety such as robustness and stability [ 6 ] . Nevertheless, these controller design techniques assume that a CBF is already given for the safety constraint considered. Constructing CBFs can be challenging, especially when dealing with safety constraints with higher relative degrees. The most popular approach for addressing this issue is High-Order CBFs (HOCBFs) [ 7 , 8 , 9 , 10 , 11 ] . While effective in some cases, HOCBFs are not traditional CBFs. Many standard results associated with CBFs, such as robustness and stability, do not readily transfer to HOCBFs, and extending these results to HOCBFs is often nontrivial [ 8 , 12 , 13 ] . Another approach to constructing CBFs is backstepping [ 14 , 15 ] , which produces CBFs rather than HOCBFs. This method requires systems to be in strict feedback form, or transformable into strict feedback form via output coordinates [ 16 ] . Backstepping involves designing a sequence of smooth virtual controllers for a sequence of auxiliary systems, which increases the complexity of control design compared to HOCBFs. While recent advancements [ 5 , 17 ] have made the design of these virtual controllers systematic, the requirements on the system’s structure may preclude the application of backstepping to more complex systems. An advantage of backstepping over HOCBFs is its ability to handle constraints with mixed-input relative degree, in the sense of independent inputs appearing at different orders of derivatives. In the context of HOCBFs, [ 18 ] addresses this issue using integral control [ 19 ] to dynamically extend inputs, materializing them at different relative degrees. While this enables controller synthesis, it obscures the original inputs in the design process, making it difficult to analyze or minimize control effort. The main contribution of this paper is the development of a method for constructing CBFs from safety constraints with higher relative degrees. Our approach extends HOCBF ideas by introducing activation functions that consider HOCBF constraints only when necessary. The result is a Rectified Control Barrier Function (ReCBF) , rather than a HOCBF, that inherits existing properties of CBFs such as stability and robustness. In addition, our approach can generate true CBFs from existing HOCBFs, and it is better suited to handle safety constraints with a weak relative degree for which HOCBF may struggle. We discuss our method by focusing first on safety constraints with relative degree two, and then we move on to mixed-input and higher relative degree constraints. Moreover, we provide a comparative analysis of ReCBFs with other methods such as HOCBF and backstepping, and illustrate how the main ideas presented herein may also be adapted to these approaches. Finally, we apply our method to a fixed-wing aircraft control problem. II Preliminaries and Problem Formulation II-A Control Barrier Functions Consider a nonlinear control affine system 1 1 1 A continuous function \alpha\,:\,(-a,b)\rightarrow\mathbb{R} , a,b\in\mathbb{R}_{>0} , is said to be an extended class \mathcal{K} function ( \alpha\in\mathcal{K}^{e} ) if \alpha(0)=0 and \alpha is strictly increasing. If a=b=\infty and \lim_{r\rightarrow\pm\infty}\alpha(r)=\pm\infty then \alpha is said to be an extended class \mathcal{K}_{\infty} function ( \alpha\in\mathcal{K}^{e}_{\infty} ). For a continuously differentiable function \alpha\,:\,\mathbb{R}\rightarrow\mathbb{R} , we define {\alpha^{\prime}(r)\coloneqq\frac{{\rm d}\alpha}{{\rm d}r}(r)} . With an abuse of terminology, we say that a function is smooth if it is differentiable as many times as necessary. For a smooth function h\,:\,\mathbb{R}^{n}\rightarrow\mathbb{R}^{p} and vector field \mathbf{f}\,:\,\mathbb{R}^{n}\rightarrow\mathbb{R}^{n} we define L_{\mathbf{f}}h(\mathbf{x})\coloneqq\frac{\partial h}{\partial\mathbf{x}}(\mathbf{x})\mathbf{f}(\mathbf{x}) as the Lie derivative of h along \mathbf{f} with higher order Lie derivatives denoted by L_{\mathbf{f}}^{i}h(\mathbf{x})\coloneqq\frac{\partial L_{\mathbf{f}}^{i-1}h}{\partial\mathbf{x}}(\mathbf{x})\mathbf{f}(\mathbf{x}) . : \dot{\mathbf{x}}=\mathbf{f}(\mathbf{x})+\mathbf{g}(\mathbf{x})\mathbf{u}, (1) with state \mathbf{x}\in\mathbb{R}^{n} and input \mathbf{u}\in\mathbb{R}^{m} , where \mathbf{f}\,:\,\mathbb{R}^{n}\rightarrow\mathbb{R}^{n} and \mathbf{g}\,:\,\mathbb{R}^{n}\rightarrow\mathbb{R}^{n\times m} are smooth functions. Given a locally Lipschitz feedback controller \mathbf{k}\,:\,\mathbb{R}^{n}\rightarrow\mathbb{R}^{m} for ( 1 ), the closed-loop system with \mathbf{u}=\mathbf{k}(\mathbf{x}) and initial condition {\mathbf{x}_{0}\in\mathbb{R}^{n}} admits a unique continuously differentiable trajectory \mathbf{x}\,:\,I(\mathbf{x}_{0})\rightarrow\mathbb{R}^{n} defined on a maximal interval of existence I(\mathbf{x}_{0})\subseteq\mathbb{R}_{\geq 0} . Our main objective in this paper is to design feedback controllers \mathbf{k} such that the closed-loop system satisfies state constraints \mathbf{x}(t)\in\mathcal{C} along trajectories, where \mathcal{C}\subset\mathbb{R}^{n} is a state constraint set. This is linked to the concept of forward invariance: a set \mathcal{S}\subset\mathbb{R}^{n} is said to be forward invariant for the closed-loop system if for each initial condition \mathbf{x}_{0}\in\mathcal{S} , the resulting trajectory satisfies \mathbf{x}(t)\in\mathcal{S} for all t\in I(\mathbf{x}_{0}) . While we may wish to design controllers that render the state constraint set \mathcal{C} forward invariant, such a controller may not exist, and one must instead search for a subset \mathcal{S}\subset\mathcal{C} that can be rendered forward invariant. A popular approach to designing controllers enforcing forward invariance of such sets is through CBFs. Definition 1 ( [ 1 ] ) . A continuously differentiable function h\,:\,\mathbb{R}^{n}\rightarrow\mathbb{R} defining a set \mathcal{S} as: \displaystyle\mathcal{S}\coloneqq \displaystyle\{\mathbf{x}\in\mathbb{R}^{n}\,:\,h(\mathbf{x})\geq 0\}, (2) is said to be a CBF for ( 1 ) on \mathcal{S}\subset\mathbb{R}^{n} if there exists \alpha\in\mathcal{K}^{e} and an open set \mathcal{E}\supset\mathcal{S} such that for all \mathbf{x}\in\mathcal{E} : \sup_{\mathbf{u}\in\mathbb{R}^{m}}\left\{L_{\mathbf{f}}h(\mathbf{x})+L_{\mathbf{g}}h(\mathbf{x})\mathbf{u}\right\}\geq-\alpha(h(\mathbf{x})). (3) The main utility of CBFs is that any locally Lipschitz controller \mathbf{k}(\cdot) satisfying ( 3 ) enforces forward invariance of \mathcal{S} [ 1 ] . In this paper, we focus on constraint sets of the form: \mathcal{C}\coloneqq\left\{\mathbf{x}\in\mathbb{R}^{n}\,:\,\psi(\mathbf{x})\geq 0\right\}, (4) where {\psi\,:\,\mathbb{R}^{n}\rightarrow\mathbb{R}} is smooth, and seek CBFs with corresponding zero superlevel sets contained within \mathcal{C}\supset\mathcal{S} . The following lemma outlines conditions for verifying CBFs. Lemma 1 ( [ 20 ] ) . A continuously differentiable function h\,:\,\mathbb{R}^{n}\rightarrow\mathbb{R} is a CBF for ( 1 ) on a set \mathcal{S} as in ( 2 ) if and only if there exists \alpha\in\mathcal{K}^{e} and an open set \mathcal{E}\supset\mathcal{S} such that: L_{\mathbf{g}}h(\mathbf{x})=\mathbf{0}\implies L_{\mathbf{f}}h(\mathbf{x})\geq-\alpha(h(\mathbf{x})),\quad\forall\mathbf{x}\in\mathcal{E}. (5) II-B High-Order Control Barrier Functions While Lemma 1 provides a simple condition for verifying a candidate CBF, proposing such a function in the first place is non-trivial for high-dimensional systems where inputs may not directly affect the safety constraint. A popular way to overcome this challenge is via HOCBFs [ 7 , 8 ] wherein a candidate CBF is dynamically extended to a new function that may serve as a certificate of safety. The success of this technique relies on the notion of relative degree. Definition 2 . A