IEEE Transactions on Signal Processing最新文献_第7页

Gohberg-Semencul Toeplitz Covariance Estimation via Autoregressive Parameters

IF 4.6 2区工程技术

IEEE Transactions on Signal Processing Pub Date : 2025-01-29 DOI: 10.1109/TSP.2025.3536101

Benedikt Böck;Dominik Semmler;Benedikt Fesl;Michael Baur;Wolfgang Utschick

{"title":"Gohberg-Semencul Toeplitz Covariance Estimation via Autoregressive Parameters","authors":"Benedikt Böck;Dominik Semmler;Benedikt Fesl;Michael Baur;Wolfgang Utschick","doi":"10.1109/TSP.2025.3536101","DOIUrl":"10.1109/TSP.2025.3536101","url":null,"abstract":"The use of prior structural knowledge is essential for the estimation of covariance matrices and their inverses when only few data samples are accessible. A well-known example is the knowledge that the covariance matrix is Toeplitz-structured, which occurs when dealing with wide-sense-stationary processes. Exploiting the close relation between autoregressive parameters and inverse covariance matrices, this paper introduces a new class of estimators for Toeplitz-structured covariance matrices and their inverses. To achieve this, we derive novel constraint sets for autoregressive parameters by leveraging their connection to the so-called Gohberg-Semencul decomposition. While these constraint sets guarantee the corresponding inverse covariance matrix to be positive definite and, thus, enable a proper estimation of the covariance matrix by inversion, they also build a means to control the estimator's performance by hyperparameter tuning. The derived constraint sets comprise simple box constraints enabling computationally cheap estimators in closed form. Due to the ensured positive definiteness, the proposed estimators perform well for both the estimation of the covariance matrix and its inverse. Extensive simulation results validate the proposed estimators’ efficacy for several standard Toeplitz-structured covariance matrices commonly employed in a wide range of applications.","PeriodicalId":13330,"journal":{"name":"IEEE Transactions on Signal Processing","volume":"73 ","pages":"858-875"},"PeriodicalIF":4.6,"publicationDate":"2025-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10857370","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143056375","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Asymptotic Error Rates for Point Process Classification

IF 4.6 2区工程技术

IEEE Transactions on Signal Processing Pub Date : 2025-01-27 DOI: 10.1109/TSP.2025.3531373

Xinhui Rong;Victor Solo

引用次数: 0

A Distributed Multi-Objective Detection Method for Multi-Sensor Systems With Unknown Local SNR

IF 4.6 2区工程技术

IEEE Transactions on Signal Processing Pub Date : 2025-01-27 DOI: 10.1109/TSP.2025.3533275

Chang Gao;Qingfu Zhang;Pramod K. Varshney;Xi Lin;Hongwei Liu

{"title":"A Distributed Multi-Objective Detection Method for Multi-Sensor Systems With Unknown Local SNR","authors":"Chang Gao;Qingfu Zhang;Pramod K. Varshney;Xi Lin;Hongwei Liu","doi":"10.1109/TSP.2025.3533275","DOIUrl":"10.1109/TSP.2025.3533275","url":null,"abstract":"Distributed detection, which fuses the preprocessed observations of the same area from local sensors, can generally improve target detection performance. For scenarios in practical applications where sensors cannot obtain the target signal-to-noise ratio (SNR) parameters in advance, non-coherent integration is mostly used for distributed detection. However, this detector is equivalent to the optimal detector only under the condition that the target SNRs of all the sensors are exactly the same. This condition is quite stringent for the observation of non-cooperative targets. This paper first compares the performance of traditional optimal detectors, the non-coherent integration (NCI) detector, and the single-sensor detector from a unified perspective based on the concept of Pareto optimality. Then, from the perspective of multi-objective optimization, the fusion rule and corresponding parameter learning method are designed. Theoretical analysis shows that the proposed non-identical SNR detection fusion rule possesses weak Pareto optimality. Simulation experiments demonstrate that the proposed method effectively achieves a trade-off between the optimal detection performance across sensors with multiple SNRs. Compared to the optimal detector in the presence of mismatch between the assumed and actual SNR of the target, the proposed method can achieve a significant improvement in detection performance. Additionally, the proposed method outperforms the NCI detector in scenarios where the SNR distributions of target observations across different sensors exhibit greater diversity.","PeriodicalId":13330,"journal":{"name":"IEEE Transactions on Signal Processing","volume":"73 ","pages":"649-663"},"PeriodicalIF":4.6,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143050363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Graph Neural Networks Over the Air for Decentralized Tasks in Wireless Networks

IF 4.6 2区工程技术

IEEE Transactions on Signal Processing Pub Date : 2025-01-27 DOI: 10.1109/TSP.2025.3534685

Zhan Gao;Deniz Gündüz

{"title":"Graph Neural Networks Over the Air for Decentralized Tasks in Wireless Networks","authors":"Zhan Gao;Deniz Gündüz","doi":"10.1109/TSP.2025.3534685","DOIUrl":"10.1109/TSP.2025.3534685","url":null,"abstract":"Graph neural networks (GNNs) model representations from networked data and allow for decentralized execution through localized communications. Existing GNNs often assume ideal communications and ignore potential channel effects, such as fading and noise, leading to performance degradation in real-world implementation. Considering a GNN implemented over nodes connected through wireless links, this paper conducts a stability analysis to study the impact of channel impairments on the performance of GNNs, and proposes graph neural networks over the air (AirGNNs), a novel GNN architecture that incorporates the communication model and permits decentralized execution with over-the-air computation. AirGNNs modify graph convolutional operations that shift graph signals over random communication graphs to account for channel fading and noise when aggregating features from neighbors, thus, improving architecture robustness to channel impairments. We develop a channel-inversion signal transmission strategy for AirGNNs when channel state information (CSI) is available, and propose a stochastic gradient descent based method to train AirGNNs when CSI is unknown. The convergence analysis shows that the training procedure approaches a stationary solution of an associated stochastic optimization problem and the variance analysis characterizes the statistical behavior of the trained model. Experiments on decentralized source localization, multi-robot flocking and wireless channel management corroborate theoretical findings and show superior performance of AirGNNs over wireless communication channels.","PeriodicalId":13330,"journal":{"name":"IEEE Transactions on Signal Processing","volume":"73 ","pages":"721-737"},"PeriodicalIF":4.6,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143050364","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An Investigation of Using Rigid Body Receivers for Locating a Non-Cooperative Object by Pseudo-Ranges in the Absence of Synchronization

IF 4.6 2区工程技术

IEEE Transactions on Signal Processing Pub Date : 2025-01-27 DOI: 10.1109/TSP.2025.3535099

Xiaochuan Ke;K. C. Ho

引用次数: 0

Subspace Constrained Variational Bayesian Inference for Structured Compressive Sensing With a Dynamic Grid

IF 4.6 2区工程技术

IEEE Transactions on Signal Processing Pub Date : 2025-01-24 DOI: 10.1109/TSP.2025.3532953

An Liu;Yufan Zhou;Wenkang Xu

{"title":"Subspace Constrained Variational Bayesian Inference for Structured Compressive Sensing With a Dynamic Grid","authors":"An Liu;Yufan Zhou;Wenkang Xu","doi":"10.1109/TSP.2025.3532953","DOIUrl":"10.1109/TSP.2025.3532953","url":null,"abstract":"We investigate the problem of recovering a structured sparse signal from a linear observation model with an uncertain dynamic grid in the sensing matrix. The state-of-the-art expectation maximization based compressed sensing (EM-CS) methods, such as turbo compressed sensing (Turbo-CS) and turbo variational Bayesian inference (Turbo-VBI), have a relatively slow convergence speed due to the double-loop iterations between the E-step and M-step. Moreover, each inner iteration in the E-step involves a high-dimensional matrix inverse in general, which is unacceptable for problems with large signal dimensions or real-time calculation requirements. Although there are some attempts to avoid the high-dimensional matrix inverse by majorization minimization, the convergence speed and accuracy are often sacrificed. To better address this problem, we propose an alternating estimation framework based on a novel subspace constrained VBI (SC-VBI) method, in which the high-dimensional matrix inverse is replaced by a low-dimensional subspace constrained matrix inverse (with the dimension equal to the sparsity level). We further prove the convergence of the SC-VBI to a stationary solution of the Kullback-Leibler divergence minimization problem. Simulations demonstrate that the proposed SC-VBI algorithm can achieve a much better tradeoff between complexity per iteration, convergence speed, and performance compared to the state-of-the-art algorithms.","PeriodicalId":13330,"journal":{"name":"IEEE Transactions on Signal Processing","volume":"73 ","pages":"781-794"},"PeriodicalIF":4.6,"publicationDate":"2025-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143030953","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

The Effectiveness of Local Updates for Decentralized Learning Under Data Heterogeneity

IF 4.6 2区工程技术

IEEE Transactions on Signal Processing Pub Date : 2025-01-24 DOI: 10.1109/TSP.2025.3533208

Tongle Wu;Zhize Li;Ying Sun

{"title":"The Effectiveness of Local Updates for Decentralized Learning Under Data Heterogeneity","authors":"Tongle Wu;Zhize Li;Ying Sun","doi":"10.1109/TSP.2025.3533208","DOIUrl":"10.1109/TSP.2025.3533208","url":null,"abstract":"We revisit two fundamental decentralized optimization methods, Decentralized Gradient Tracking (DGT) and Decentralized Gradient Descent (DGD), with multiple local updates. We consider two settings and demonstrate that incorporating local update steps can reduce communication complexity. Specifically, for <inline-formula><tex-math>$mu$</tex-math></inline-formula>-strongly convex and <inline-formula><tex-math>$L$</tex-math></inline-formula>-smooth loss functions, we proved that local DGT achieves communication complexity <inline-formula><tex-math>$tilde{mathcal{O}}Big{(}frac{L}{mu(K+1)}+frac{delta+{}{mu}}{mu(1-rho)}+frac{rho}{(1-rho)^{2}}cdotfrac{L+delta}{mu}Big{)}$</tex-math></inline-formula>, where <inline-formula><tex-math>$K$</tex-math></inline-formula> is the number of additional local update, <inline-formula><tex-math>$rho$</tex-math></inline-formula> measures the network connectivity and <inline-formula><tex-math>$delta$</tex-math></inline-formula> measures the second-order heterogeneity of the local losses. Our results reveal the tradeoff between communication and computation and show increasing <inline-formula><tex-math>$K$</tex-math></inline-formula> can effectively reduce communication costs when the data heterogeneity is low and the network is well-connected. We then consider the over-parameterization regime where the local losses share the same minimums. We proved that employing local updates in DGD, even without gradient correction, achieves exact linear convergence under the Polyak-Łojasiewicz (PL) condition, which can yield a similar effect as DGT in reducing communication complexity. Customization of the result to linear models is further provided, with improved rate expression. Numerical experiments validate our theoretical results.","PeriodicalId":13330,"journal":{"name":"IEEE Transactions on Signal Processing","volume":"73 ","pages":"751-765"},"PeriodicalIF":4.6,"publicationDate":"2025-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143030952","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Wasserstein Distributionally Robust Graph Learning via Algorithm Unrolling

IF 4.6 2区工程技术

IEEE Transactions on Signal Processing Pub Date : 2025-01-23 DOI: 10.1109/TSP.2025.3526287

Xiang Zhang;Yinfei Xu;Mingjie Shao;Yonina C. Eldar

{"title":"Wasserstein Distributionally Robust Graph Learning via Algorithm Unrolling","authors":"Xiang Zhang;Yinfei Xu;Mingjie Shao;Yonina C. Eldar","doi":"10.1109/TSP.2025.3526287","DOIUrl":"10.1109/TSP.2025.3526287","url":null,"abstract":"In this paper, we consider inferring the underlying graph topology from smooth graph signals. Most existing approaches learn graphs by minimizing a well-designed empirical risk using the observed data, which may be prone to data uncertainty that arises from noisy measurements and limited observability. Therefore, the learned graphs may be unreliable and exhibit poor out-of-sample performance. To enhance the robustness to data uncertainty, we propose a smoothness-based graph learning framework from a distributionally robust perspective, which is equivalent to solving an <inline-formula><tex-math>$mathrm{inf-sup}$</tex-math></inline-formula> problem. However, learning graphs directly in this way is challenging since (i) the <inline-formula><tex-math>$mathrm{inf-sup}$</tex-math></inline-formula> problem is intractable, and (ii) many parameters need to be manually determined. To address these issues, we first reformulate the <inline-formula><tex-math>$mathrm{inf-sup}$</tex-math></inline-formula> problem into a tractable one, where robustness is achieved via a regularizer. Theoretically, we show that the regularizer can improve generalization of the proposed graph estimator by bounding the out-of-sample risks. We then propose an algorithm based on the ADMM framework to solve the induced problem and further unroll it into a neural network. All parameters are determined automatically and simultaneously by training the unrolled network. Extensive experiments on both synthetic and real-world data demonstrate that our approach can achieve superior and more robust performance than existing models on different observed signals.","PeriodicalId":13330,"journal":{"name":"IEEE Transactions on Signal Processing","volume":"73 ","pages":"676-690"},"PeriodicalIF":4.6,"publicationDate":"2025-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143026443","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

On Optimal MMSE Channel Estimation for One-Bit Quantized MIMO Systems 论一位量化多输入多输出系统的最优 MMSE 信道估计

IF 4.6 2区工程技术

IEEE Transactions on Signal Processing Pub Date : 2025-01-21 DOI: 10.1109/TSP.2025.3531779

Minhua Ding;Italo Atzeni;Antti Tölli;A. Lee Swindlehurst

{"title":"On Optimal MMSE Channel Estimation for One-Bit Quantized MIMO Systems","authors":"Minhua Ding;Italo Atzeni;Antti Tölli;A. Lee Swindlehurst","doi":"10.1109/TSP.2025.3531779","DOIUrl":"10.1109/TSP.2025.3531779","url":null,"abstract":"This paper focuses on the minimum mean squared error (MMSE) channel estimator for multiple-input multiple-output (MIMO) systems with one-bit quantization at the receiver side. Despite its optimality and significance in estimation theory, the MMSE estimator has not been fully investigated in this context due to its general nonlinearity and computational complexity. Instead, the typically suboptimal Bussgang linear MMSE (BLMMSE) channel estimator has been widely adopted. In this work, we develop a new framework to compute the MMSE channel estimator that hinges on the computation of the orthant probability of a multivariate normal distribution. Based on this framework, we determine a necessary and sufficient condition for the BLMMSE channel estimator to be optimal and thus equivalent to the MMSE estimator. Under the assumption of specific channel correlation or pilot symbols, we further utilize the framework to derive analytical expressions for the MMSE estimator that are particularly convenient for the computation when certain system dimensions become large, thereby enabling a comparison between the BLMMSE and MMSE channel estimators in these cases.","PeriodicalId":13330,"journal":{"name":"IEEE Transactions on Signal Processing","volume":"73 ","pages":"617-632"},"PeriodicalIF":4.6,"publicationDate":"2025-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10848316","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142992794","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Dynamic Spectrum Cartography: Reconstructing Spatial-Spectral-Temporal Radio Frequency Map via Tensor Completion 动态频谱制图：通过张量补全重构空间-光谱-时间无线电频率图

IF 4.6 2区工程技术

IEEE Transactions on Signal Processing Pub Date : 2025-01-21 DOI: 10.1109/TSP.2025.3531872

Xiaonan Chen;Jun Wang;Qingyang Huang

{"title":"Dynamic Spectrum Cartography: Reconstructing Spatial-Spectral-Temporal Radio Frequency Map via Tensor Completion","authors":"Xiaonan Chen;Jun Wang;Qingyang Huang","doi":"10.1109/TSP.2025.3531872","DOIUrl":"10.1109/TSP.2025.3531872","url":null,"abstract":"Spectrum cartography (SC) aims to construct a global radio-frequency (RF) map across multiple domains, e.g., space, frequency and time, from sparse sensor samples. Recent state-of-the-art SC methods have successfully established the recoverability of <inline-formula><tex-math>$3$</tex-math></inline-formula>-D spatial-spectral RF maps using identifiable models, such as non-negative matrix factorization (NMF) and block-term tensor decomposition (BTD). However, these models do not account for possible time dynamics in RF environment. This work takes a step forward and focuses on a <inline-formula><tex-math>$4$</tex-math></inline-formula>-D spatial-spectral-temporal SC task under time-varying scenarios. From a data recovery viewpoint, the task is highly ill-posed since the degree of freedom (DoF) in a <inline-formula><tex-math>$4$</tex-math></inline-formula>-D map is extremely high. To address this issue, a two-stage methodology is put forth: for stage one, sensor measurements are unraveled into incomplete RF map w.r.t each emitter; for stage two, individual RF maps are completed in parallel and then synthesize the <inline-formula><tex-math>$4$</tex-math></inline-formula>-D map. In this way, DoF in the recovery process is significantly reduced. Two different algorithms are designed, including a basic batch-based one and a full-fledged streaming one enabling on-line SC. From the theory side, recoverability of the proposed approaches is characterized by certain sampling patterns or complexity. Experiments using synthetic, ray-tracing, and real-world data are employed to showcase the effectiveness of the proposed methods.","PeriodicalId":13330,"journal":{"name":"IEEE Transactions on Signal Processing","volume":"73 ","pages":"1184-1199"},"PeriodicalIF":4.6,"publicationDate":"2025-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142992785","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0