Framelet perturbation and application to nouniform sampling approximation for Sobolev space

TL;DR

This paper develops a framelet-based method for approximating functions in Sobolev spaces using nonuniform samples, demonstrating exponential convergence and robustness to sample jittering.

Contribution

It introduces a novel framelet approximation operator for Sobolev spaces that achieves exponential convergence and stability under nonuniform sampling perturbations.

Findings

Exponential convergence rate of the framelet approximation in Sobolev spaces.

Robustness of the approximation operator to shift perturbations and jittering.

Establishment of nonuniform sampling approximation for all functions in Sobolev spaces.

Abstract

The Sobolev space $H^{s}(\mathbb{R}^{d})$, where $s > d/2$, is an important function space that has many applications in various areas of research. Attributed to the inertia of a measuring instrument, it is desirable in sampling theory to reconstruct a function by its nonuniform samples. In the present paper, we investigate the problem of constructing the approximation to all the functions in $H^{s}(\mathbb{R}^{d})$ with nonuniform samples by utilizing dual framelet systems for the Sobolev space pair $(H^{s}(\mathbb{R}^{d}), H^{-s}(\mathbb{R}^{d}))$. We first establish the convergence rates of the framelet series in $(H^{s}(\mathbb{R}^{d}), H^{-s}(\mathbb{R}^{d}))$, and then construct the framelet approximation operator holding for the entire space $H^{s}(\mathbb{R}^{d})$. Using the approximation operator, any function in $H^{s}(\mathbb{R}^{d})$ can be approximated at the exponential rate with respect to the scale level. We examine the stability property for the perturbations of the framelet approximation operator with respect to shift parameters, and obtain an estimate bound for the perturbation error. Our result shows that under the condition $s > d/2$, the approximation operator is robust to the shift perturbation. These results are used to establish the nonuniform sampling approximation for every function in $H^{s}(\mathbb{R}^{d})$. In particular, the new nonuniform sampling approximation error is robust to the jittering of the samples.