Multi-dimensional unlimited sampling and robust reconstruction

From MaRDI portal





The paper introduces a new framework for sampling and reconstructing multi-dimensional analog signals, extending the Unlimited Sensing Framework (USF). The authors define a new folded sampling operator called the multi-dimensional modulo-hysteresis, which generalizes the classical one-dimensional modulo operator and is designed to exploit the inherent redundancy of multi-dimensional data.\N\NThey prove that the proposed operator is well-defined and that its outputs have a bounded dynamic range. For the noiseless case, the paper establishes theoretical guarantees for exact input reconstruction. In the presence of Gaussian noise, the method uses redundancy in the \(D-1\) remaining dimensions to reduce error probability, showing that this probability tends to \(0\) as the sampling rate increases.\N\NFinally, the authors present numerical experiments that support the theoretical results and demonstrate the robustness of the proposed approach to higher noise levels compared to the standard USF model.











This page was built for publication: Multi-dimensional unlimited sampling and robust reconstruction

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6932281)