What problem does this paper attempt to address?

This paper attempts to solve the problem of multichannel spectral super - resolution, especially in the case where the signal has a constant amplitude (CA) or a non - constant amplitude. Specifically, the paper focuses on estimating frequencies in multiple channels from partially observed data. ### Problem Background Multichannel spectral super - resolution refers to estimating the frequencies of individual complex sinusoidal waves from superimposed samples of multiple channels. The sinusoidal waves in each channel share the same frequency but may have different phases and amplitudes. When only partially observed data are available, how to efficiently and accurately recover these frequencies is a challenge. ### Research Motivation Classical methods such as MUSIC and ESPRIT can achieve arbitrarily high resolution when dealing with complete noiseless samples, but their performance will decline significantly in the case of incomplete observations. To overcome this limitation, researchers have proposed methods based on sparse optimization and compressed sensing, especially Atomic Norm Minimization (ANM). ANM deals with incomplete observations by transforming the super - resolution problem into a semidefinite programming (SDP) problem, but its high computational complexity limits its application in large - scale problems. ### Paper Contributions To solve the above problems, the author proposes two optimization problems based on low - rank Hankel - Toeplitz matrix decomposition and designs corresponding low - complexity gradient - descent algorithms. Specifically: 1. **General Multichannel Signals**: For multichannel signals with non - constant amplitudes, the author proposes a new optimization problem and eliminates rank and positive - semidefinite constraints through low - rank Hankel - Toeplitz matrix decomposition. 2. **Constant - Amplitude Signals**: For multichannel signals with constant amplitudes, the author also proposes a similar optimization problem and designs a corresponding low - complexity algorithm. ### Main Results - **Theoretical Analysis**: It is proved that the newly proposed optimization problems are equivalent to the traditional Hankel - Toeplitz matrix optimization problems. - **Algorithm Design**: Two low - complexity gradient - descent algorithms (MHTGD and CHTGD) are proposed, and the gradients are efficiently calculated through the Fast Fourier Transform (FFT). - **Numerical Experiments**: The effectiveness and superiority of the proposed algorithms are verified through a large number of numerical experiments, especially in terms of computational efficiency, which is significantly better than existing methods such as ANM and SACA. ### Conclusion By introducing low - rank Hankel - Toeplitz matrix decomposition, this paper successfully reduces the computational complexity of the multichannel spectral super - resolution problem and provides an efficient algorithm implementation. The experimental results show that the proposed algorithms are comparable to existing methods in performance but have obvious advantages in computational efficiency. ### Related Formulas - Multichannel Signal Model: \[ x^\star_{jl}=\sum_{k = 1}^{K}b^\star_{kl}e^{i(- 2\pi f^\star_k\cdot(j - 1)+\phi^\star_{kl})},\quad j\in\mathcal{N},l\in\mathcal{L} \] where \(\mathcal{N}\triangleq\{1,\ldots,N\}\), \(\mathcal{L}\triangleq\{1,\ldots,L\}\), \(N\) is the total number of samples in each channel, \(L\) is the number of channels, \(K\) is the number of frequencies, \(f^\star_k\in\mathcal{F}\triangleq[0,1)\), \(b^\star_{kl}>0\), \(\phi^\star_{kl}\in\mathbb{R}\) respectively represent the \(k\) - th unknown normalized frequency, amplitude and phase in the \(l\) - th channel. - Incomplete Observation Model: \[ P_\Omega(X^\star)=\sum_{j\in\Omega}e_j^T X^\star

Low-Complexity Algorithms for Multichannel Spectral Super-Resolution

Multichannel Frequency Estimation with Constant Amplitude via Convex Structured Low-Rank Approximation

Simultaneous Blind Demixing and Super-resolution via Vectorized Hankel Lift

Hyperspectral Image Super-Resolution Using Sparse Spectral Unmixing and Low-Rank Constraints

Deep Nonlocal Low-Rank Regularization for Complex-Domain Pixel Super-Resolution.

An Efficient Algorithm for Video Super-Resolution Based On a Sequential Model

Robust Superresolution of Multiangle-Multispectral Remote Sensing Images Based on Rank Minimization.

Spatial-Spectral Structured Sparse Low-Rank Representation for Hyperspectral Image Super-Resolution

Fast and Provable Simultaneous Blind Super-Resolution and Demixing for Point Source Signals: Scaled Gradient Descent without Regularization

Connecting Low-Level and High-Level Visions: A Joint Optimization for Hyperspectral Image Super-Resolution and Target Detection

Simpler Gradient Methods for Blind Super-Resolution with Lower Iteration Complexity

Spectral CT reconstruction via low-rank representation and structure preserving regularization

Blind Super-Resolution of Point Sources Via Fast Iterative Hard Thresholding

How Does the Low-Rank Matrix Decomposition Help Internal and External Learnings for Super-Resolution

Hyperspectral super-resolution via low rank tensor triple decomposition

An Efficient Algorithm for Single-/Multi-Channel Sar Superresolution Imaging of Large Scenes

Low Complexity Single Image Super-Resolution with Channel Splitting and Fusion Network

Hybrid-Field Channel Estimation for XL-MIMO: A Proximal Gradient Algorithm on the Fixed-Rank Matrix Manifold

Vectorized Hankel Lift: A Convex Approach for Blind Super-Resolution of Point Sources

Multichannel deconvolution with a high-frequency structural regularization

Hyperspectral Super-Resolution: A Coupled Tensor Factorization Approach