Frames for source recovery from non-uniform dynamical samples
Ruchi, Lalit Kumar Vashisht · 2025
Read Frames for source recovery from non-uniform dynamical samples by Ruchi, Lalit Kumar Vashisht (2025) 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-2501-16949.txt.
Introduction
Ruchi and Lalit Kumar Vashisht study how to recover a hidden source from samples of a moving system in a 2025 mathematics study. The system evolves in steps over a non-uniform set of times that arises from spectral pairs, inside an infinite-dimensional space. Earlier work by Aldroubi and colleagues frames the recovery problem and its stability, which frames must guarantee.
The authors give a condition that is both necessary and sufficient for recovering the source in finitely many steps. They also derive a necessary condition for stable recovery when the source lies in a closed subspace. Stable recovery matters for sensors that watch dynamics indirectly.
Catalog record
| CREATOR | Ruchi, Lalit Kumar Vashisht |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 96 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-16949
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Frames for source recovery from non-uniform dynamical samples Ruchi, Lalit Kumar Vashisht Abstract Motivated by the work of Aldroubi et al., we investigate the stability of the source term of the discrete dynamical system indexing over a non-uniform discrete set arising from spectral pairs in infinite-dimensional separable Hilbert spaces. Extending results due to Aldroubi et al., firstly, we give a necessary and sufficient condition for the recovery of the source term in finitely many iterations. Afterwards, we derive a necessary condition for the stability of the source term in finitely many iterations when it belongs to the closed subspace of an infinite-dimensional separable Hilbert space. Finally, we give a necessary and sufficient condition for the recovery of the source term in infinitely many iterations. AMS Subject Classification (2020): 93B30, 42C15, 42C30. Keywords: Sampling theory; forcing; frames; reconstruction. 1 Introduction The interplay between frame theory and dynamical sampling has emerged as a riveting area of research in functional analysis, offering profound insights into signal processing, data science, and applied mathematics. Frames were pioneered by Duffin and Schaeffer in a fundamental published paper [ 17 ] in 1952 while studying some deep problems in nonharmonic Fourier series. In a nutshell, they endeavored to find condition(s) under which a countable family of complex exponentials in the signal space L^{2}(-\gamma,\gamma) , \gamma>0 , form a Riesz basis or are complete in L^{2}(-\gamma,\gamma) . The mathematical definition of frames in separable Hilbert spaces is given in Sec. 2 of the paper. Frames were reviewed in a book by Young [ 32 ] . For fundamental properties of frames and their applications in pure mathematics and engineering science, we refer to texts [ 21 , 22 , 24 , 28 ] . In one of the directions of applications of frames in engineering science, dynamical sampling is a relatively recent framework that addresses the challenge of reconstructing signals from spatially and temporally evolving measurements. The origins of dynamical sampling trace back to the seminal work of Aldroubi et al. [ 2 , 3 ] in 2013, where they investigated conditions under which a function can be reconstructed from coarse samples of its iterates under a linear operator. Recent research has also accentuated applying dynamical sampling to new challenges, such as real-time data streaming, dynamic environments, and non-stationary signals. These advancements have permeated the applicability of dynamical sampling techniques in areas ranging from medical diagnostics and environmental monitoring to robotics and artificial intelligence. Aldroubi et al. studied applications of Bessel sequences, frames, and Riesz bases in sampling theory in a series of papers, see [ 1 , 2 , 4 , 7 ] and many references therein. On the other hand, Fuglede [ 18 ] in the study of commuting self-adjoint partial differential operators introduced the concept of spectral pairs. By using the theory of spectral pairs in the Lebesgue space L^{2}(\mathbb{R}) , Gabardo and Nashed [ 19 ] generalized the notion of multiresolution analysis (MRA). They introduced the concept of non-uniform multiresolution analysis. In non-uniform multiresolution analysis, the corresponding translation set may not be a group but is a spectrum associated with one-dimensional spectral pairs. It is applicable in input/output models based on shifts of data points, where shifts may not be uniform in nature. Frame conditions for the discrete Gabor system related to spectral pairs can be found in [ 31 ] . 1.1 Related work and contribution Dynamical sampling problems can be stratified in three types depending on the specific variable of interest. When the objective is to determine the conditions on spatial sampling set for recovering initial state, the problem is referred to as the space-time trade-off [ 2 , 6 , 7 , 9 , 14 , 16 ] . On the other hand, when the aim is to determine the conditions for the recovery of dynamic operator A , the problem is categorized as system identification [ 5 , 8 , 15 , 30 ] and the problem of identifying specific types of source terms that drive the dynamics of the system is known as source recovery problems , see [ 10 , 11 , 12 , 13 ] and many references therein. The formal development of dynamical sampling as a distinct area began in the early 2010s. In 2015, Aldroubi et al. introduced the concept of sampling in evolving systems in their landmark work in [ 3 ] . They studied sampling and reconstruction for functions evolving under the action of linear operators, laying the groundwork for future studies. Aldroubi, Cabrelli, and Molter [ 7 ] developed the initial framework for finite-dimensional spaces and certain classes of self-adjoint operators in infinite-dimensional spaces, providing the theoretical foundation for subsequent studies. In [ 14 ] , Cabrelli, Molter and Petrosyan extended this work to investigate dynamical sampling on finite index sets. They provided a theoretical solution to the problem where a signal evolving over time through iterates of an operator can be recovered from its space-time samples. By 2021, the focus shifted towards the reconstruction of bandlimited functions from space–time samples. Researchers analyzed the problem of reconstructing a bandlimited function from the space–time samples of its states resulting from convolution with a kernel, addressing challenges in stable reconstruction [ 9 ] . In recent work, Huang, Neuman, Tang, and Xie explored convolutional dynamical sampling, focusing on the recovery of bandlimited heat diffusion fields from space-time samples in both finite and infinite-dimensional cases. Their work extended previous findings and emphasized the density of space-time sampling sets [ 23 ] . For applications of sampling in sensor networks, it is necessary to mention the published papers [ 25 , 29 ] . Most recently, Aldroubi et al. [ 13 ] advanced the field by characterizing frames for stable recovery of source terms in dynamical systems. Specifically, they consider the discrete dynamical system of the form x_{n+1}=Ax_{n}+w,n\in\mathbb{N},w\in W, where x_{n}\in\mathcal{H} is the n -th state of the system, and \mathcal{H} is a separable Hilbert space. The operator A\in\mathcal{B}(\mathcal{H}),x_{0}\in\mathcal{H} is the initial state, and W is a closed subspace of \mathcal{H} . Time-space sample measurements \displaystyle D(x_{0},w)=[\langle x_{n},g_{j}\rangle]_{n\geq 0,\,j\geq 1} are obtained by inner products \langle x_{n},g_{j}\rangle with vectors of a Bessel system \{g_{j}\}_{j\geq 1}\subset\mathcal{H} . It is given in [ 13 ] that frame conditions related to the above dynamical system can be used in environmental monitoring applications for identifying the locations and magnitude of pollution sources. Aldroubi et al., in [ 13 ] , focused on the stable recovery of a source term by using time-space sample measurements formed by inner products with vectors from a given Bessel system in separable Hilbert spaces. They proved necessary and sufficient conditions for the stable recovery of constant source terms from time-space samples in separable Hilbert spaces. On the other hand, Gabardo and Nashed, in [ 19 ] , generalized the concept of multiresolution analysis (MRA) by using the theory of spectral pairs in the Lebesgue space L^{2}(\mathbb{R}) . The theory of spectral pairs and non-uniform wavelets is instrumental in developing advanced numerical methods for solving partial differential equations, particularly in cases where traditional grid-based techniques fail or are computationally expensive. In this framework, the translation set is not necessarily a group but is instead a spectrum