Coordinates adapted to vector fields. III: Real analyticity (Q2236428)

The paper under review is the third from a series of papers dealing with a special coordinate system adapted to a set of vector fields of class \(C^1\) on a \(C^2\)-manifold, which at every point generates the tangent space. The results in these papers are related to the conditions under which the considered vector fields are smooth, or real analytic, or have Zygmund regularity of some finite order. In the first paper from the series [\textit{B. Stovall} and \textit{B. Street}, Geom. Funct. Anal. 28, No. 6, 1780--1862 (2018; Zbl 1443.58004)], the authors provide a diffeomorphism invariant version of some results in [\textit{A. Nagel} et al., Acta Math. 155, 103--147 (1985; Zbl 0578.32044)] on the quantitative theory of sub-Riemannian (Carnot-Caratheodory) geometry. The results are one derivative short of being optimal. Optimal regularity results in terms of Zgymund spaces are provided in the second paper [\textit{B. Street}, Am. J. Math. 143, No. 6, 1791--1840 (2021; Zbl 1489.58002)], by using methods of elliptic PDE's. Since this approach is not applicable to the real analytic setting, in the paper under review, the third of the series, the author uses ODE technics to obtain real analyticity results. By taking the vector fields \(X^1, X^2, \dots,X^q\) that span the tangent space at every point of a \(C^2\)-manifold \(M\), the author provides necessary and sufficient conditions under which there exists a coordinate chart near a fixed point of \(M\), such that \(X^1, X^2, \dots,X^q\) are real analytic in that coordinate system. Next, he obtains equivalent conditions for the existence of a unique real analytic manifold structure on \(M\), compatible with its \(C^2\)-structure, such that the vector fields are real analytic with respect to this structure. To apply these qualitative results in analysis, the author shows how the coordinate chart can be picked so that \(X^1, X^2, \dots,X^q\) are normalized with respect to it, and proves that this type of coordinate systems can be seen as scaling maps in sub-Riemannian geometry, thus continuing the results obtained in [\textit{A. Nagel} et al., Acta Math. 155, 103--147 (1985; Zbl 0578.32044); \textit{C. L. Fefferman} and \textit{A. Sánchez-Calle}, Ann. Math. (2) 124, 247--272 (1986; Zbl 0613.35002); \textit{T. Tao} and \textit{J. Wright}, J. Am. Math. Soc. 16, No. 3, 605--638 (2003; Zbl 1080.42007); \textit{B. Street}, Rev. Mat. Iberoam. 27, No. 2, 645--732 (2011; Zbl 1222.53036)]. The author applies some of the main results of the paper on the special background of the Euclidean space with Lebesgue measure.