Detection of Jointly Sparse Signals via Locally Most Powerful Tests with Generalized Gaussian Model

Xueqian Wang

2023-01-01

Abstract:In Chap. 2 , we consider the problem of detection of jointly sparse signals with Gaussian signals and noise.

What problem does this paper attempt to address?

Detection of Jointly Sparse Signals via Locally Most Powerful Tests with Gaussian Noise

Xueqian Wang

2023-01-01

Abstract:In recent years, compressive sensing (CS) has emerged as a new paradigm for sparse signal processing, which aims at obtaining valuable information of sparse signals from a small number of measurements.
Detection of Sparse Signals in Sensor Networks Via Locally Most Powerful Tests

Xueqian Wang,Gang Li,Pramod K. Varshney

DOI: https://doi.org/10.1109/lsp.2018.2861222

2018-01-01

IEEE Signal Processing Letters

Abstract:We consider the problem of detection of sparse stochastic signals with a distributed sensor network. Multiple sensors in the network are assumed to observe sparse signals, which share the joint sparsity pattern. The Bernoulli-Gaussian (BG) distribution with sparsity-enforcing capability is imposed on the sparse signals. The sparsity degree in the BG model is positive and close to zero in the presence of the sparse signals and is zero in the absence of the signals. Motivated by this, the problem of detection of the sparse signals with a distributed sensor network is formulated as the problem of close and one-sided hypothesis testing on the sparsity degree. For this problem, we propose a detector based on the locally most powerful test (LMPT) to decide on the presence or absence of sparse signals with sensor networks. The proposed LMPT detector does not require signal recovery, which alleviates the complexity of the detection system in sensor networks. Simulation results illustrate the performance of the proposed LMPT detector and corroborate our theoretical analysis. Simulation results also show that, compared to the detector based on matching pursuit, the proposed LMPT detector significantly reduces the computational burden without noticeable performance loss.
Distributed Detection of Generalized Gaussian Sparse Signals with One-Bit Measurements (Poster).

Xueqian Wang,Gang Li,Chen Quan,Pramod K. Varshney

DOI: https://doi.org/10.23919/fusion43075.2019.9011411

2019-01-01

Abstract:In this paper, distributed detection of sparse stochastic signals with one-bit measurements is studied. We assume that both the noise and the dominant elements in sparse signals follow the generalized Gaussian (GG) distribution. Due to constrained bandwidth/energy in sensor networks, the local sensors send binary measurements instead of analog data to the fusion center. First, we propose the locally most powerful test (LMPT) detector based on one-bit measurements for distributed detection of sparse signals. Then, we analytically derive near-optimal one-bit quantizers at the local sensor nodes using the log-concave approximation of the efficacy. Simulation results corroborate our theoretical analysis and show that, the proposed one-bit LMPT detector with analytically obtained one-bit quantizers provides detection performance which is comparable to that with numerically obtained optimal quantizers in the GG case.
Distributed Detection of Sparse Stochastic Signals with Quantized Measurements: the Generalized Gaussian Case

Xueqian Wang,Gang Li,Chen Quan,Pramod K. Varshney

DOI: https://doi.org/10.1109/tsp.2019.2932884

IF: 4.875

2019-01-01

IEEE Transactions on Signal Processing

Abstract:In this paper, we consider distributed detection of sparse stochastic signals with quantized measurements. Assume that both the noise and the dominant elements in sparse signals follow the generalized Gaussian (GG) distribution. Since the communication bandwidth in sensor networks is limited, the local sensors send quantized measurements instead of analog data to the fusion center. First, we propose the locally most powerful test (LMPT) detector based on quantized measurements for distributed detection of sparse signals with the GG distribution. Then, the local quantizers are designed by maximizing the efficacy of the proposed quantized LMPT detector. In particular, we obtain the near-optimal closed-form solutions of 1-bit quantizers at the local sensors using the log-concave approximation of efficacy. Multilevel quantizers at the local sensors are numerically obtained using the particle swarm optimization (PSO) algorithm. The asymptotic relative efficiency (ARE) is derived analytically to measure the performance loss of the proposed LMPT detector caused by quantization of local measurements. Simulation results corroborate our theoretical analysis of the proposed quantized LMPT detector.
Adaptive Double Subspace Signal Detection in Gaussian Background—Part II: Partially Homogeneous Environments

Weijian Liu,Wenchong Xie,Jun Liu,Yongliang Wang

DOI: https://doi.org/10.1109/tsp.2014.2309553

IF: 4.875

2014-05-01

IEEE Transactions on Signal Processing

Abstract:In this part of the paper, we continue to study the problem of detecting a double subspace signal in Gaussian noise. Precisely, we address the detection problem in partially homogeneous environments, where the primary and secondary data share the same covariance matrix up to an unknown scaling factor. We derive the generalized likelihood ratio test (GLRT), Rao test, Wald test, and their two-step versions. We also introduce three spectral norm tests (SNTs). All these detectors possess the constant false alarm rate (CFAR) property. Moreover, various kinds of special cases of these detectors are given. At the stage of performance evaluation, we consider two cases. One is the case of no signal mismatch. The other is more general, namely, the case of signal mismatch, including the column-space signal mismatch and row-space signal mismatch.

engineering, electrical & electronic
Multichannel Signal Detection Based on Wald Test in Subspace Interference and Gaussian Noise.

Weijian Liu,Jun Liu,Hai Li,Qinglei Du,Yong-Liang Wang

DOI: https://doi.org/10.1109/taes.2018.2870445

IF: 3.491

2018-01-01

IEEE Transactions on Aerospace and Electronic Systems

Abstract:This paper investigates the problem of detecting a multichannel signal embedded in subspace interference and Gaussian noise. The interference lies in a known subspace but with unknown coordinates, while the noise, in the general sense, consisting of thermal noise and clutter, has an unknown covariance matrix. To estimate the covariance matrix, it is assumed that there are sufficient signal-free training data. Two cases are considered, namely, the homogeneous environment (HE) and partially HE (PHE). The test and training data in the HE have the same noise covariance matrix, while the test and training data in the PHE share the same noise covariance matrix up to an unknown scaling factor. We derive the corresponding Wald tests both in the HE and PHE. Remarkably, these two detectors reject the clutter by using the sample covariance matrix, and then reject interference and integrate signal simultaneously through the oblique projection. Numerical examples show that the two proposed detectors can provide better detection performance than existing detectors in some situations.
Adaptive Double Subspace Signal Detection in Gaussian Background—Part I: Homogeneous Environments

Weijian Liu,Wenchong Xie,Jun Liu,Yongliang Wang

DOI: https://doi.org/10.1109/tsp.2014.2309556

IF: 4.875

2014-05-01

IEEE Transactions on Signal Processing

Abstract:In this two-part paper, we consider the problem of adaptive multidimensional/multichannel signal detection in Gaussian noise with unknown covariance matrix. The test data (primary data) is assumed as a collection of sample vectors, arranged as the columns of a rectangular data array. The rows and columns of the signal matrix are both assumed to lie in known subspaces, but with unknown coordinates. Due to this feature of the signal structure, we name this kind of signal as the double subspace signal. Part I of this paper focuses on the adaptive detection in homogeneous environments, while Part II deals with the adaptive detection in partially homogeneous environments. Precisely, in this part, we derive the generalized likelihood ratio test (GLRT), Rao test, Wald test, as well as their two-step variations, in homogeneous environments. Three types of spectral norm tests (SNTs) are also introduced. All these detectors are shown to possess the constant false alarm rate (CFAR) property. Moreover, we discuss the differences between them and show how they work. Another contribution is that we investigate various special cases of these detectors. Remarkably, some of them are well-known existing detectors, while some others are still new. At the stage of performance evaluation, conducted by Monte Carlo simulations, both matched and mismatched signals are dealt with. For each case, more than one scenario is considered.

engineering, electrical & electronic
Generalized spectral testing for multivariate continuous-time models

Bin Chen,Yongmiao Hong

DOI: https://doi.org/10.1016/j.jeconom.2011.06.001

IF: 3.363

2011-01-01

Journal of Econometrics

Abstract:We develop an omnibus specification test for multivariate continuous-time models using the conditional characteristic function, which often has a convenient closed-form or can be accurately approximated for many multivariate continuous-time models in finance and economics. The proposed test fully exploits the information in the joint conditional distribution of underlying economic processes and hence is expected to have good power in a multivariate context. A class of easy-to-interpret diagnostic procedures is supplemented to gauge possible sources of model misspecification. Our tests are also applicable to discrete-time distribution models. Simulation studies show that the tests provide reliable inference in finite samples.
Persymmetric Detection of Subspace Signals Embedded in Subspace Interference and Gaussian Noise

Jun Liu,Jiajun Li,Weijian Liu

DOI: https://doi.org/10.23919/eusipco47968.2020.9287666

2021-01-01

Abstract:We consider the problem of detecting subspace signals embedded in subspace interference and Gaussian noise with unknown covariance matrix. According to the criterion of generalized likelihood ratio test, we exploit persymmetry to propose an adaptive detector. Moreover, the statistical characterization of the proposed detector in the absence of target signals is obtained, which exhibits a constant false alarm rate property against the noise covariance matrix. Numerical examples illustrate that the proposed detector outperforms its counterparts, especially when the number of training data is small.
Adaptive Robust Subspace Detection Based on GLRT, Rao, Wald, Gradient, and Durbin Tests

Gaoqing Xiong,Hui Cao,Weijian Liu,Jun Liu,Chongying Qi,Daikun Zheng

DOI: https://doi.org/10.1007/s00034-024-02740-z

IF: 2.311

2024-01-01

Circuits Systems and Signal Processing

Abstract:The current paper investigates the issue of designing adaptive robust subspace detectors in Gaussian noise whose covariance matrix is unknown. The original problem is revised by importing a fictitious signal with a given structure within the signal-plus-noise hypothesis to collect leakage signals around the subspace, thus increasing the credibility of this hypothesis in situations involving mismatch. To solve the issue described above, we utilize the generalized likelihood ratio test, Rao, Wald, Gradient, and Durbin tests to derive five adaptive subspace detectors. Both theoretical proofs and Monte Carlo simulation results suggest that these proposed detectors possess the constant false alarm rate properties. Numerical examples reveal the effectiveness of these proposed detectors and show their varying degrees of robustness under mismatch scenarios.
A Support Detection and Root Finding Approach for Learning High-dimensional Generalized Linear Models

Jian Huang,Yuling Jiao,Lican Kang,Jin Liu,Yanyan Liu,Xiliang Lu

DOI: https://doi.org/10.48550/arxiv.2001.05819

2020-01-01

Abstract:Feature selection is important for modeling high-dimensional data, where the number of variables can be much larger than the sample size. In this paper, we develop a support detection and root finding procedure to learn the high dimensional sparse generalized linear models and denote this method by GSDAR. Based on the KKT condition for $\ell_0$-penalized maximum likelihood estimations, GSDAR generates a sequence of estimators iteratively. Under some restricted invertibility conditions on the maximum likelihood function and sparsity assumption on the target coefficients, the errors of the proposed estimate decays exponentially to the optimal order. Moreover, the oracle estimator can be recovered if the target signal is stronger than the detectable level. We conduct simulations and real data analysis to illustrate the advantages of our proposed method over several existing methods, including Lasso and MCP.
Multiuser Detection with Sparse Spectrum Gaussian Process Regression

Shaowei Wang,Hualai Gu

DOI: https://doi.org/10.1109/lcomm.2011.120211.111508

IF: 3.5529

2012-01-01

IEEE Communications Letters

Abstract:Multiuser detection in direct-sequence code-division multiple access (DS-CDMA) systems can be implemented by using Gaussian process (GP) for regression in the sense of minimum mean square error criterion. In this Letter we investigate the application of sparse spectrum Gaussian process (SSGP) to the multiuser detection problem. The key point of the SSGP is that the sparsity of spectral representation of Gaussian process leads to an algorithm with much lower complexity than the full GP, while keeping almost the same bit error rate (BER) for DS-CDMA systems. Experimental results validate our proposed SSGP based multiuser detection method. It achieves greater efficiency and good BER performance.
Global Testing under the Sparse Alternatives for Single Index Models

Qian Lin,Zhigen Zhao,Jun S. Liu

DOI: https://doi.org/10.48550/arxiv.1805.01820

2018-01-01

Abstract:For the single index model $y=f(\beta^{\tau}x,\epsilon)$ with Gaussian design, %satisfying that rank $var(\mathbb{E}[x\mid y])=1$ where $f$ is unknown and $\beta$ is a sparse $p$-dimensional unit vector with at most $s$ nonzero entries, we are interested in testing the null hypothesis that $\beta$, when viewed as a whole vector, is zero against the alternative that some entries of $\beta$ is nonzero. Assuming that $var(\mathbb{E}[x \mid y])$ is non-vanishing, we define the generalized signal-to-noise ratio (gSNR) $\lambda$ of the model as the unique non-zero eigenvalue of $var(\mathbb{E}[x \mid y])$. We show that if $s^{2}\log^2(p)\wedge p$ is of a smaller order of $n$, denoted as $s^{2}\log^2(p)\wedge p\prec n$, where $n$ is the sample size, one can detect the existence of signals if and only if gSNR$\succ\frac{p^{1/2}}{n}\wedge \frac{s\log(p)}{n}$. Furthermore, if the noise is additive (i.e., $y=f(\beta^{\tau}x)+\epsilon$), one can detect the existence of the signal if and only if gSNR$\succ\frac{p^{1/2}}{n}\wedge \frac{s\log(p)}{n} \wedge \frac{1}{\sqrt{n}}$. It is rather surprising that the detection boundary for the single index model with additive noise matches that for linear regression models. These results pave the road for thorough theoretical analysis of single/multiple index models in high dimensions.
Minimax Signal Detection in Sparse Additive Models

Subhodh Kotekal,Chao Gao

2024-10-02

Abstract:Sparse additive models are an attractive choice in circumstances calling for modelling flexibility in the face of high dimensionality. We study the signal detection problem and establish the minimax separation rate for the detection of a sparse additive signal. Our result is nonasymptotic and applicable to the general case where the univariate component functions belong to a generic reproducing kernel Hilbert space. Unlike the estimation theory, the minimax separation rate reveals a nontrivial interaction between sparsity and the choice of function space. We also investigate adaptation to sparsity and establish an adaptive testing rate for a generic function space; adaptation is possible in some spaces while others impose an unavoidable cost. Finally, adaptation to both sparsity and smoothness is studied in the setting of Sobolev space, and we correct some existing claims in the literature.

Statistics Theory,Methodology,Machine Learning
On Minimax Robust Detection of Stationary Gaussian Signals in White Gaussian Noise

Wenyi Zhang,H. V. Poor

DOI: https://doi.org/10.1109/TIT.2011.2136210

IF: 2.5

2011-01-01

IEEE Transactions on Information Theory

Abstract:The problem of detecting a wide-sense stationary Gaussian signal process embedded in white Gaussian noise, in which the power spectral density of the signal process exhibits uncertainty, is investigated. The performance of minimax robust detection is characterized by the exponential decay rate of the miss probability under a Neyman-Pearson criterion with a fixed false alarm probability, as the length of the observation interval grows without bound. A stochastic suppression condition is identified for the uncertainty set of spectral density functions, and it is established that, under the stochastic suppression condition, the resulting minimax problem possesses a saddle point, which is achievable by the likelihood ratio tests matched to a so-called suppressing power spectral density in the uncertainty set. No convexity condition on the uncertainty set is required to establish this result.
Persymmetric Rao And Wald Tests For Adaptive Detection Of Distributed Targets In Compoundgaussian Noise

Xiaolu Guo,Haihong Tao,Hong-Yan Zhao,Jun Liu

DOI: https://doi.org/10.1049/iet-rsn.2016.0251

2017-01-01

Abstract:The problem of detecting a distributed target in the presence of compound-Gaussian noise with unknown covariance matrix is studied in this paper. Since no uniformly most powerful test exists for the problem at hand, two detectors based on the Rao and Wald tests are devised. Remarkably, the persymmetric structure of the covariance matrix is exploited in the design of the proposed detectors. Simulation results show that the proposed detectors outperform the traditional detectors, especially in training-limited scenarios.
A Unified Recovery of Structured Signals Using Atomic Norm

Xuemei Chen

2023-11-11

Abstract:In many applications we seek to recover signals from linear measurements far fewer than the ambient dimension, given the signals have exploitable structures such as sparse vectors or low rank matrices. In this paper we work in a general setting where signals are approximately sparse in an so called atomic set. We provide general recovery results stating that a convex programming can stably and robustly recover signals if the null space of the sensing map satisfies certain properties. Moreover, we argue that such null space property can be satisfied with high probability if each measurement is subgaussian even when the number of measurements are very few. Some new results for recovering signals sparse in a frame, and recovering low rank matrices are also derived as a result.

Information Theory,Functional Analysis
Sequential joint signal detection and signal-to-noise ratio estimation

M. Fauß,K. G. Nagananda,A. M. Zoubir,H. V. Poor

DOI: https://doi.org/10.48550/arXiv.1701.05297

2017-01-19

Abstract:The sequential analysis of the problem of joint signal detection and signal-to-noise ratio (SNR) estimation for a linear Gaussian observation model is considered. The problem is posed as an optimization setup where the goal is to minimize the number of samples required to achieve the desired (i) type I and type II error probabilities and (ii) mean squared error performance. This optimization problem is reduced to a more tractable formulation by transforming the observed signal and noise sequences to a single sequence of Bernoulli random variables; joint detection and estimation is then performed on the Bernoulli sequence. This transformation renders the problem easily solvable, and results in a computationally simpler sufficient statistic compared to the one based on the (untransformed) observation sequences. Experimental results demonstrate the advantages of the proposed method, making it feasible for applications having strict constraints on data storage and computation.

Information Theory
Minimax detection boundary and sharp optimal test for Gaussian graphical models

Yumou Qiu,Bin Guo

DOI: https://doi.org/10.1093/jrsssb/qkae029

2024-03-30

Abstract:Abstract In this article, we derive the minimax detection boundary for testing a sub-block of variables in a precision matrix under the Gaussian distribution. Compared to the results on the minimum rate of signals for testing precision matrices in literature, our result gives the exact minimum signal strength in a precision matrix that can be detected. We propose a thresholding test that is able to achieve the minimax detection boundary under certain cases by adaptively choosing the threshold level. The asymptotic distribution of the thresholding statistic for precision matrices is derived. Power analysis is conducted to show the proposed test is powerful against sparse and weak signals, which cannot be detected by the existing Lmax and L2 tests. Simulation studies show the proposed test has an accurate size around the nominal level and is more powerful than the existing tests for detecting sparse and weak signals in precision matrices. Real data analysis on brain imaging data is carried out to illustrate the utility of the proposed test in practice, which reveals functional connectivity between brain regions for Alzheimer’s disease patients and normal healthy people.

statistics & probability

Detection of Jointly Sparse Signals via Locally Most Powerful Tests with Generalized Gaussian Model

Detection of Jointly Sparse Signals via Locally Most Powerful Tests with Gaussian Noise

Detection of Sparse Signals in Sensor Networks Via Locally Most Powerful Tests

Distributed Detection of Generalized Gaussian Sparse Signals with One-Bit Measurements (Poster).

Distributed Detection of Sparse Stochastic Signals with Quantized Measurements: the Generalized Gaussian Case

Adaptive Double Subspace Signal Detection in Gaussian Background—Part II: Partially Homogeneous Environments

Multichannel Signal Detection Based on Wald Test in Subspace Interference and Gaussian Noise.

Adaptive Double Subspace Signal Detection in Gaussian Background—Part I: Homogeneous Environments

Generalized spectral testing for multivariate continuous-time models

Persymmetric Detection of Subspace Signals Embedded in Subspace Interference and Gaussian Noise

Adaptive Robust Subspace Detection Based on GLRT, Rao, Wald, Gradient, and Durbin Tests

A Support Detection and Root Finding Approach for Learning High-dimensional Generalized Linear Models

Multiuser Detection with Sparse Spectrum Gaussian Process Regression

Global Testing under the Sparse Alternatives for Single Index Models

Minimax Signal Detection in Sparse Additive Models

On Minimax Robust Detection of Stationary Gaussian Signals in White Gaussian Noise

Persymmetric Rao And Wald Tests For Adaptive Detection Of Distributed Targets In Compoundgaussian Noise

A Unified Recovery of Structured Signals Using Atomic Norm

Sequential joint signal detection and signal-to-noise ratio estimation

Minimax detection boundary and sharp optimal test for Gaussian graphical models