Quantitative versions of Pansu Asymptotic Theorem and of Mitchell Tangent Theorem
Enrico Le Donne, Sebastiano Nicolussi Golo, Andrea Tettamanti
TL;DR
This work provides quantitative refinements of two foundational geometric limits for geodesic Lie groups: the Pansu asymptotic cone in nilpotent, sub-Finsler settings and the Mitchell tangent cone in general sub-Finsler settings. By introducing asymptotic and tangent gradings, along with constants α_infty, α_0 and β, the authors derive explicit power-law rates for |d(p,q)−d_infty(p,q)| and |d(p,q)−d_0(p,q)| on compacts, improving prior sublinear bounds. The methodology hinges on one-parameter families of contracted and dilated metrics, Grönwall-type lemmas adapted to each setting, and the notion of Carnot quotient ideals to localize the comparison to Carnot factors; this yields uniform convergence results and explicit GH convergence rates. The results illuminate how choices of gradings affect convergence speed and provide sharper bounds in natural constructions (e.g., products with Carnot factors), with clear implications for the study of sub-Finsler geometry and its infinitesimal limits.
Abstract
We quantitatively study the speed of convergence of geodesic Lie groups to their metric limits. For nilpotent geodesic Lie groups, we give estimates on the difference of the original metrics and the asymptotic metrics, while for general geodesic Lie groups, we give similar estimates for the difference of the original metrics and the tangent metrics. In both settings, our results sharpen existing bounds in the literature.
