Manifolds of positive reach, differentiability, tangent variation, and attaining the reach

2024/12/06 by André Lieutier, Lieutier, André, Mathijs Wintraecken +1 · 1 citation
Engineering · #Robotic Mechanisms and Dynamics

paper · pdf · doi:10.48550/arxiv.2412.04906

Abstract

This paper contains three main results. Firstly, we give an elementary proof of the following statement: Let M be a (closed, in both the geometrical and topological sense of the word) topological manifold embedded in ℝ^d. If M has positive reach, then M can locally be written as the graph of a C^1,1 from the tangent space to the normal space. Conversely if M can locally written as the graph of a C^1,1 function from the tangent space to the normal space, then M has positive reach. The result was hinted at by Federer when he introduced the reach, and proved by Lytchak. Lytchak's proof relies heavily CAT(k)-theory. The proof presented here uses only basic results on homology. Secondly, we give optimal Lipschitz-constants for the derivative, in other words we give an optimal bound for the angle between tangent spaces in term of the distance between the points. This improves earlier results, that were either suboptimal or assumed that the manifold was C². Thirdly, we generalize a result by Aamari et al. which explains the how the reach is attained from the smooth setting to general sets of positive reach.

Cited by

• High-Dimensional Statistical Process Control via Manifold Fitting and Learning

Related

• A free lunch: manifolds of positive reach can be smoothed without decreasing the reach
• A note on subsets of positive reach
• Lower bounds for the reach and applications
• Tight bounds for the learning of homotopy à la Niyogi, Smale, and Weinberger for subsets of Euclidean spaces and of Riemannian manifolds
• Universally consistent estimation of the reach