Complex Sensing: Reconstructing Dense Rational Signals From Few Complex Measurements
IEEE Signal Processing Letters · 2026 · 33 · 3098-3101
DOI: 10.1109/lsp.2026.3714224 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von William H. Pan und 2 weiteren Autor:innen. Die Arbeit nennt 10 Literaturangaben. Gefördert von: NSF.
Abstract (Originalsprache)
Reconstructing high-dimensional signals from few linear measurements is a classic problem in signal processing. When the target signal is sparse, the broad compressed sensing literature provides sufficient conditions over the sensing matrix so that reconstruction is possible. In this paper it is shown that if the signal’s entries are rational, and the measurement coefficients are complex, then reconstructing the signal from a few measurements is possible even without assuming sparsity. Our results make use of fundamental independence properties of roots of unity over the rationals. We construct explicit sensing matrices using such roots of unity by judiciously selecting rows from a Hadamard matrix, and show that our construction is optimal with respect to the interplay between the number of rows (i.e., measurements) and the order of the roots of unity. In practice, our results imply that low-resolution (e.g., integer) signals can be reconstructed from few high-resolution measurements even without being sparse. We formally prove the feasibility of this type of reconstruction, provide explicit and optimal sensing matrices, formulate reconstruction algorithms using known lattice algorithms, and demonstrate good experimental performance.
Förderung
- NSF
Zitieren
APA
Pan, W. H., Roscow, D., & Raviv, N. (2026). Complex Sensing: Reconstructing Dense Rational Signals From Few Complex Measurements. IEEE Signal Processing Letters, 33, 3098-3101. https://doi.org/10.1109/lsp.2026.3714224
Vancouver
Pan WH, Roscow D, Raviv N. Complex Sensing: Reconstructing Dense Rational Signals From Few Complex Measurements. IEEE Signal Processing Letters. 2026;33:3098-3101. doi:10.1109/lsp.2026.3714224
BibTeX
@article{Pan2026,
author = {Pan, William H. and Roscow, Dylan and Raviv, Netanel},
title = {Complex Sensing: Reconstructing Dense Rational Signals From Few Complex Measurements},
journal = {IEEE Signal Processing Letters},
year = {2026},
volume = {33},
pages = {3098--3101},
doi = {10.1109/lsp.2026.3714224}
}
EUVIMED