Unlimited sensing of multiband signals
File(s)
Author(s)
Shtendel, Gal
Type
Thesis or dissertation
Abstract
Analog-to-digital converters (ADCs) underpin modern information systems, where a signal's time and amplitude are quantized. While time quantization can be lossless for certain signal models, amplitude quantization is a bottleneck that limits dynamic range and resolution. The Unlimited Sensing Framework (USF) is a sampling paradigm that reconstructs signals from the quantization error rather than from quantized amplitudes. This leads to high dynamic range (HDR) and high digital resolution (HDRes) acquisition with fewer bits. This thesis develops the USF for multiband signals, which are sums of narrowband components across a wide spectrum, common in radar, communications, and audio applications. In multiband settings, HDR is crucial to capture both weak and strong signal components, while HDRes is essential for accurate recovery from sub-Nyquist measurements, due to the wide span of the spectral support. Conceptually, the goal is to exploit the Fourier structure of multiband signals within the USF. However, this is nontrivial: the nonlinear mechanism that enables HDR also distorts the waveform, so the observed sequence is neither bandlimited nor exhibits a simple spectral pattern. This thesis addresses this challenge through theory, algorithm design, and hardware validation. It analyses single-and multi-channel architectures, establishes conditions under which a multiband structure enables stable sub-Nyquist recovery in the USF, and derives corresponding reconstruction methods. Hardware experiments corroborate the theory and demonstrate practical feasibility. Collectively, the results show that the USF can exploit Fourier structure to enable lowpower, HDR, sub-Nyquist sensing architectures. A key advantage is improved e ective resolution without increasing bit depth, which is particularly valuable for sub-Nyquist schemes, where fewer measurements demand high-resolution ADCs. The ndings have implications for radar, communications, and related systems, bridging the gap between sampling theory and deployable solutions.
Version
Open Access
Date Issued
2025-10-29
Date Awarded
2026-07-01
Copyright Statement
Attribution-NonCommercial 4.0 International Licence (CC BY-NC)
License URL
Advisor
Bhandari, Ayush
Publisher Department
Department of Electrical and Electronic Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)