Search

Two‐dimensional (2D) direction‐of‐arrival (DOA) estimation is crucial in array signal processing. Compressed sensing (CS) provides a superior alternative to spatial spectrum estimation algorithms by enabling 2D DOA estimation of correlated sources from single snapshot data. However, the grid mismatch effect inherent in grid‐based CS algorithms impacts estimation accuracy. Despite recent advancements, the state‐of‐the‐art gridless CS algorithm, decoupled atomic norm minimization, is limited to specific 2D array geometries, such as uniform rectangular arrays. This letter presents an efficient gridless 2D DOA estimation algorithm for generalized rectangular arrays, including both uniform and sparse arrays. The proposed algorithm achieves high accuracy through a novel approach called generalized matrix‐form atomic norm minimization and provides a fast solution using the alternating direction method of multipliers. Validation through computer simulations and practical experiments underscores its efficacy. [ABSTRACT FROM AUTHOR]

Greedy pursuit algorithms are widely used in sparse signal processing for their computational efficiency. However, research on the reconstruction error properties is far from comprehensive. This paper derives the statistical properties of reconstruction error in greedy pursuit algorithms, including probability density function (PDF), expectation, and covariance. The reconstruction error follows a mixture distribution, which is composed of multiple multivariate random variables and weighted by the probability of support sets. The multivariate random variable obeys a truncated distribution that results from restricting the noise domain. The validity of the derivations is verified by using the orthogonal matching pursuit (OMP) algorithm as an example. [ABSTRACT FROM AUTHOR]

This paper solves the problem of block sparse vector recovery using the block ℓ1−αℓq$\ell _1-\alpha \ell _q$‐minimization model. Based on the block restricted isometry property (B‐RIP) condition, exact block sparse vector recovery result is obtained. The theoretical bound for the block ℓ1−αℓq$\ell _1-\alpha \ell _q$‐minimization model are also obtained when measurements are depraved by the noises. [ABSTRACT FROM AUTHOR]

This paper, considers the channel estimation problem of the uplink low Earth orbit (LEO) satellite orthogonal frequency division multiplexing (OFDM) system. By exploiting the channel structured sparsity in the delay‐Doppler‐angle domain, the channel estimation problem is formulated as a block‐sparse signal recovery problem. The variance state propagation (VSP) framework is extended to solve such a problem. Considering that the delay, Doppler shift, and angle of departure (AoD) of each propagation path may not exactly fall on the presumed discrete grid points, an expectation maximization (EM) based parameter learning scheme in VSP to iteratively update the grid parameters is introduced. Simulation results demonstrate that the proposed approach achieves a substantial performance gain over the existing methods. [ABSTRACT FROM AUTHOR]

In this paper, the application of a hybrid representation in wide‐angle synthetic aperture radar (WASAR) imaging is investigated, addressing the challenge of achieving sparse representation in the presence of complex electromagnetic scattering characteristics and highly anisotropic targets. A convolutional neural network is utilized to represent the two‐dimensional data within each azimuth aperture, while employing blind compressed sensing (BCS) to achieve sparse representation across different azimuth apertures. Convolutional neural network (CNN) excels in learning spatial hierarchies and local dependencies of two‐dimensional data but requires a large amount of training data. Isotropic targets within each azimuth aperture can be used for training conventional SAR data, while acquiring training samples for anisotropic targets poses challenges. To address this issue, BCS are integrated into WASAR imaging, eliminating the need for additional training data. By integrating these methods, a novel approach called hybrid‐WASAR is proposed, which incorporates two regularization terms into WASAR imaging and employs the alternating direction method of multipliers to iteratively solve the imaging model. Compared to traditional WASAR imaging techniques, hybrid‐WASAR improves the accuracy of reconstructing the backscattering coefficients of the targets, leading to a significant enhancement in overall imaging quality. [ABSTRACT FROM AUTHOR]

To complete the sparse sampling of linear frequency modulation (LFM) signals and estimate the parameters, the authors proposed an improved finite rate of innovation (FRI) system to achieve the goal. First the authors build a mathematical model based on the Z‐transform, then they input the LFM signal into the modified FRI system and process the output of the system based on this mathematical model, which enables the estimation of the parameters of the LFM signal using only four sampling points. The error of the final experimental results can be controlled within 3%, which greatly improves the efficiency of LFM signal parameter estimation. [ABSTRACT FROM AUTHOR]

This paper focuses on the compressed sensing ℓ1−ℓ2$\ell _1-\ell _2$‐minimization model and develops new bounds on cumulative coherence μ1(s)$\mu _1(s)$. It is pointed out that if cumulative coherence μ1(s)$\mu _1(s)$ satisfies Equation (2) or (11), then the sparse signal can stably recover in noise model and exactly recover in free noise by ℓ1−ℓ2$\ell _1-\ell _2$‐minimization model. From this paper, it is found that based on some condition of cumulative coherence, the ℓ1−ℓ2$\ell _1-\ell _2$‐minimization model can exactly recover s‐sparse signals in noiseless cases and stably recover s‐sparse signals in the noise cases. [ABSTRACT FROM AUTHOR]

This letter develops a fast spectral peak estimation method, which is inspired by the theory of compressed sensing (CS). In this work, the measurement procedure of the CS is employed to construct an underdetermined equation, and then a low complexity spectral peak search method is established by using a priori knowledge of signal amplitude. Compared with other frequency estimation methods, the proposed method requires only 10% inner products demanded to obtain the spectral peak positions, and this method can be introduced into any existing discrete Fourier transform (DFT)‐based interpolation methods. Experimental results show that the proposed method yields nearly the same accuracy as other interpolation methods, but with lower computational demands. [ABSTRACT FROM AUTHOR]

In the field of compressed sensing, ℓ1−2$\ell _{1-2}$‐minimization model can recover the sparse signal well. In dealing with the ℓ1−2$\ell _{1-2}$‐minimization problem, most of the existing literature uses the difference of convex algorithm (DCA) to solve the unrestricted ℓ1−2$\ell _{1-2}$‐minimization model, that is, model (4). Although experiments have proved that the unrestricted ℓ1−2$\ell _{1-2}$‐minimization model can recover the original sparse signal, the theoretical proof has not been established yet. This paper mainly proves theoretically that the unrestricted ℓ1−2$\ell _{1-2}$‐minimization model can recover the sparse signal well, and makes an experimental study on the parameter λ in the unrestricted minimization model. The experimental results show that increasing the size of parameter λ in (4) appropriately can improve the recovery success rate. However, when λ is sufficiently large, increasing λ will not increase the recovery success rate. [ABSTRACT FROM AUTHOR]

This letter proposes a novel time delay estimation method based on sparse optimisation of the cross‐correlation function to improve the estimation accuracy of delay parameters of chirp signals in multipath environments. In this method, the time delay estimation model is converted into a correlation‐function based model to estimate the parameter of exponent signals with frequency information, and the covariance matrix of such correlation function is solved by a sparse iteration optimisation algorithm abbreviated as CCF‐SPICE. Experimental simulations show that the proposed delay estimation algorithm has a sharper peak with less power leakage compared with the existing SPICE algorithm. Its performance is better than the SPICE algorithm and conventional MUSIC algorithm, especially under low SNR in multipath environments. The MSE performance of the proposed CCF‐SPICE algorithm is closer to Cramer–Rao bounds. [ABSTRACT FROM AUTHOR]

Most existing Bayesian compressive sensing (BCS) algorithms are developed in real numbers. This results in many difficulties in applying BCS to solve complex‐valued problems. To overcome this limitation, this letter extends the existing real‐valued BCS framework to the complex‐valued BCS framework. Within this framework, the multitask learning setting, where L tasks are statistically interrelated and share the same prior, is considered. It is verified by numerical examples that the developed complex multitask compressive sensing (CMCS) algorithm is more accurate and effective than the existing algorithms for the complex sparse signal reconstructions [ABSTRACT FROM AUTHOR]

When there is the coexistence of uncorrelated and coherent signals in sparse arrays, the conventional algorithms using coarray are fail. In order to solve this problem, the letter proposes a novel method based on compressed sensing. Firstly, the authors vectorize the covariance matrix and establish a sparse representation model through constructing a two‐dimensional redundant dictionary. Then, the authors use an improved orthogonal matching pursuit algorithm for off‐grid sources to recover the sparse vector. Through analysing location of non‐zero elements in sparse vector, the direction‐of‐arrivals of both uncorrelated and coherent signals can be obtained. The proposed method has no strict limitation by the structure of the existing sparse arrays. Moreover, it makes full use of vectorized data and can estimate more number of signals than that of sensors. Numerical experiments prove the effectiveness and favourable performance of the proposed method. [ABSTRACT FROM AUTHOR]

Compressive sensing (CS) is an effective algorithm for reconstructing images from a small sample of data. CS models combining traditional optimisation‐based CS methods and deep learning have been used to improve image reconstruction performance. However, if the sample ratio is very low, the performance of the CS method combined with deep learning will be unsatisfactory. In this letter, a deep learning‐based CS model incorporating hierarchical knowledge distillation to improve image reconstruction even at varied sample ratios. Compared to the state‐of‐art methods with all compressive sensing ratios, the proposed method improved performance by an average of 0.26 dB without additional trainable parameters. [ABSTRACT FROM AUTHOR]

In this Letter, the authors propose using compressive sensing (CS)‐based Stolt migration imaging approach to reduce the amount of measurement data and provide an enhanced imaging quality. The authors utilise both forward and backward Stolt migration imaging operators to deal with the measurement and image the space data in two‐dimensional matrix form for sparse recovery computation. This imaging process accounts for the electromagnetic wave propagation and does not require prior knowledge of the transmitted pulse, which is practically not available in practice. Results of the proposed numerical simulation demonstrate the validity and effectiveness of the proposed imaging algorithm. [ABSTRACT FROM AUTHOR]

Spectral imaging technique plays a very vital role in the field of chemical detection and identification. Conventional spectroscopic imaging techniques suffer from massive acquisition time. This limitation sometimes restricts it from many practical applications. The acquisition of a full spectral image requires huge acquisition time. In this Letter, a compressive sensing-based single-pixel camera architecture has been realised to acquire spectral images that can be used for non-destructive testing and classification of explosive materials. The compressive measurements for all the spectral images are done simultaneously thus reducing the acquisition time significantly. The spectro-spatial images were reconstructed using the basis pursuit algorithm and compared with least square solutions, which resulted in fast acquisition and improved image quality. The maximum compression rate achieved was 95.84%. [ABSTRACT FROM AUTHOR]

In this study, the authors mainly discuss the robust signal recovery by ℓ1−2 minimisation with incorporating prior support information, which is not considered in previous works. A robust recovery condition is established and an recovery error estimation is obtained, in particular, the obtained results generalise the state‐of‐the‐art ones. In addition, by proposing a modified algorithm, the numerical experiments show that incorporating prior support information for ℓ1−2 minimisation exhibits better recovery performance than standard ℓ1−2 minimisation. [ABSTRACT FROM AUTHOR]

The authors propose a novel method for significantly reducing antenna measurement time and increasing the efficiency of collecting the data which most is relevant information. They apply methods from compressed sensing which reduce redundancy in the antenna measurement, i.e. a metasurface is added to the antenna near‐field, and the unit cell is manipulated by a FPGA to facilitate sparse sampling. In comparison with the traditional sparse sampling method, experimental results demonstrate that their strategy can reduce the measuring time by at least 56% while satisfying measurement error and that it has considerable potential for antenna measurement. [ABSTRACT FROM AUTHOR]

Compressive 2D near-field millimetre-wave (MMW) imaging is a compressed sensing-based technique, which reconstructs an image from a few under-sampled measurements. Although this technique can extremely improve imaging process efficiency, resolution of the reconstructed image is still limited by the scanning grid interval, which is always hard to be improved in practice due to hardware constraint. To reconstruct a higher resolution image from under-sampled measurements, a compressive 2D near-field MMW super-resolution (SR) imaging model is established. Meanwhile, the corresponding algorithm, which is based on the primal-dual framework, is also proposed. Experimental results show that the proposed algorithm can effectively reconstruct the SR MMW image in superior performance. [ABSTRACT FROM AUTHOR]

In this study, the authors mainly discuss the robust signal recovery by ℓ1−2 minimisation with incorporating prior support information, which is not considered in previous works. A robust recovery condition is established and an recovery error estimation is obtained, in particular, the obtained results generalise the state-of-the-art ones. In addition, by proposing a modified algorithm, the numerical experiments show that incorporating prior support information for ℓ1−2 minimisation exhibits better recovery performance than standard ℓ1−2 minimisation. [ABSTRACT FROM AUTHOR]

In this letter, the authors propose block normalised iterative hard thresholding (BNIHT) algorithm for the recovery of block sparse signal, in which the non-zero elements are presented in clusters. Based on block restricted isometry property, the sufficient conditions to guarantee the convergence of BNIHT are derived. In addition, the number of required iterations is obtained. The simulation experiment shows that BNIHT algorithm is superior to the block IHT (BIHT) algorithm when the step size satisfies μ < 1. [ABSTRACT FROM AUTHOR]

Block orthogonal matching pursuit is an efficient reconstruction algorithm in compressed sensing, which exploits block sparsity during support index selection. In this letter, to further improve the performance, the authors propose two block sparse reconstruction algorithms by incorporating the prior information of block support probability. Based on Gamma distribution approximation, such information is formulated as an additive term during index selection. Moreover, the second algorithm extends the first one to the scenario with inaccurate prior information by introducing an additional judging mechanism with block correlation and prior factor simultaneously. Numerical results show that the proposed algorithms outperform existing algorithms. [ABSTRACT FROM AUTHOR]

Block orthogonal matching pursuit is an efficient reconstruction algorithm in compressed sensing, which exploits block sparsity during support index selection. In this letter, to further improve the performance, the authors propose two block sparse reconstruction algorithms by incorporating the prior information of block support probability. Based on Gamma distribution approximation, such information is formulated as an additive term during index selection. Moreover, the second algorithm extends the first one to the scenario with inaccurate prior information by introducing an additional judging mechanism with block correlation and prior factor simultaneously. Numerical results show that the proposed algorithms outperform existing algorithms. [ABSTRACT FROM AUTHOR]

In this letter, the authors propose block normalised iterative hard thresholding (BNIHT) algorithm for the recovery of block sparse signal, in which the non‐zero elements are presented in clusters. Based on block restricted isometry property, the sufficient conditions to guarantee the convergence of BNIHT are derived. In addition, the number of required iterations is obtained. The simulation experiment shows that BNIHT algorithm is superior to the block IHT (BIHT) algorithm when the step size satisfies μ<1. [ABSTRACT FROM AUTHOR]

The authors propose a novel method for significantly reducing antenna measurement time and increasing the efficiency of collecting the data which most is relevant information. They apply methods from compressed sensing which reduce redundancy in the antenna measurement, i.e. a metasurface is added to the antenna near-field, and the unit cell is manipulated by a FPGA to facilitate sparse sampling. In comparison with the traditional sparse sampling method, experimental results demonstrate that their strategy can reduce the measuring time by at least 56% while satisfying measurement error and that it has considerable potential for antenna measurement. [ABSTRACT FROM AUTHOR]

The sensor selection is a technique to reduce the cost, energy consumption, and complexity of a system by discarding redundant or less useful sensors. In this technique, one attempts to select a subset of sensors within a larger set so as to optimise a performance criterion. In this work, the authors consider the detection of sources in the 3D space when the source space could be sparsely represented. The application of compressive sensing (CS) methods in this setting has been extensively studied. Their aim in this work is to carry out the task of sensor selection in a planar array without sacrificing the detection performance. Their approach is to adopt the equivalent CS model and reduce the mutual coherence of the associated sensing matrix. This minimisation task is non-convex, for which they propose a convex relaxation. Numerical simulations demonstrate that the gap between their relaxed solution and its optimal counterpart is negligible. The proposed sensor selection method is also shown to outperform the uniform approach in terms of source detection. [ABSTRACT FROM AUTHOR]

In this Letter, the authors propose a novel revised regularisation to improve the performance of compressive sensing (CS) reconstruction. They suppose that a specific regularisation term is insufficient to accommodate the prior information of CS while it can be improved by further imposing a linear approximation term. They also prove that the revised regularisation is substantially equivalent to the CS preprocessing methods. They conduct extensive experiments on various CS algorithms, which show the effectiveness of their revised regularisation. [ABSTRACT FROM AUTHOR]

It has been shown that more accurate signal recovery can be achieved with low‐coherence dictionaries in sparse signal processing. In this Letter, the authors extend the low‐coherence attribute to receptive fields in convolutional neural networks. A new constrained formulation to train low‐coherence convolutional neural network is presented and an efficient algorithm is proposed to train the network. The resulting formulation produces a direct link between the receptive fields of a layer through training procedure that can be used to extract more informative representations from the subsequent layers. Simulation results over three benchmark datasets confirm superiority of the proposed low‐coherence convolutional neural network over the unconstrained version. [ABSTRACT FROM AUTHOR]

Generalised space shift keying (GSSK) technique proposed for massive multi‐input–multi‐output systems has a higher information transmission rate due to the activation of multiple antennas at the same time, and also has significant advantage in terms of hardware cost. However, the maximum likelihood detector has a very high complexity which makes it computationally intractable for large‐scale GSSK systems. In this Letter, a sparse detector is proposed by exploiting the inherent property of sparsity in GSSK. Different from the existing compressed sensing (CS)‐based detectors, the proposed detector utilises Euclidean distance instead of inner product operation for antenna matching. Moreover, it can remove erroneous antenna indices by backtracking to promote the detection performance. The simulation results show that the proposed scheme outperforms the existing CS‐based detectors while maintaining low complexity. [ABSTRACT FROM AUTHOR]

A carrier frequency estimation algorithm for the real signal based on the modulated wideband converter compressed sampling (CS) wideband digital receiver without signal reconstruction is proposed. First, the weighted coefficients matrix can be obtained by mixing function in the receiver. Then, between every two branches of the CS receiver, the results of CS signals multiplied by the inverse of their corresponding weighted coefficients sub‐matrix are all equal. According to the equivalent relationship above, the weighted coefficients sub‐matrix corresponding received signal can be searched in a weighted coefficients matrix. In addition, the weighted coefficients sub‐matrix corresponding received signal includes the information of signal‐band index which represents the frequency support. Finally, the carrier frequency estimation of the received signal can be obtained by signal‐band index and frequency estimation of CS signals. The simulation results demonstrate the efficiency of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Nested array (NA) has received much attention for providing more degrees of freedom compared with the traditional uniform linear array. An off‐grid compressed sensing (CS) method is proposed using atomic norm minimisation for NA to identify the unknown direction‐of‐arrival of multiple sources, which overcomes the drawback of grid‐mismatch effect of the conventional CS algorithms such as orthogonal matching pursuit, and achieves better performance than the subspace methods. Simulation results verify the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

In spread spectrum (SS) communications, the input signal is spread over a wider bandwidth to avoid interference or interception. One of the most common SS techniques is direct sequence SS (DSSS). In this Letter, the authors propose non‐cooperative compressive detection techniques for DSSS signals. The proposed compressive detection framework allows the use of random as well as designed measurement kernels. A technique for designing compressive measurement kernels which exploit DSSS signal structure and non‐uniform usage of spreading codes is proposed. Theoretical and simulation results are provided to compare the performance of the proposed methods with their conventional counterparts. [ABSTRACT FROM AUTHOR]

The block orthogonal matching pursuit (BOMP) algorithm is an efficient method in compressed sensing (CS) for the reconstruction of block‐sparse signals, whose non‐zero entries occur in clusters. However, due to the non‐ideal factors in practice, there exits perturbation in the CS system, which may cause significant performance degradation during reconstruction. In this Letter, a perturbed BOMP algorithm is presented to deal with this problem, which extends BOMP algorithm to the perturbation case. The proposed algorithm performs controlled perturbation on each selected block of support vectors to decrease the orthogonal residual at each iteration. Moreover, the condition of successful reconstruction is derived. Simulation results demonstrate the effectiveness and robustness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

A semi‐definite programming‐total least squares (SDP‐TLS) method based on compressed sensing for the off‐grid direction‐of‐arrival (DOA) estimation with gain/phase uncertainties is proposed. On a discrete grid, the accuracy of DOA estimation will degrade because of the basis mismatch when the true DOAs do not coincide with the angular grid that is divided. The accuracy of DOA estimation would be worse if the gain/phase uncertainties are present but overlooked. In order to eliminate the influence of the gain/phase uncertainties and solve the off‐grid problem at the same time, a SDP‐TLS method is proposed. First, the problem needed to be solved into a framework of total least‐squares (TLS) is transformed. Then an alternating descent algorithm to solve this problem is developed. In the proposed algorithm, the off‐grid DOA is estimated by semi‐definite programming, using TLS approach to update gain/phase uncertainties through iterations. Simulation results show that the SDP‐TLS method can help to improve the accuracy of off‐grid DOA estimation with the gain/phase uncertainties than the existing algorithms. [ABSTRACT FROM AUTHOR]

Combining the best‐first tree search and orthogonal matching pursuit (OMP), A*OMP has been introduced that aims at solving the compressed sensing problem. The theoretical analysis of A*OMP is mainly presented and an improved condition for exact recovery of sparse signals via the A*OMP algorithm is proposed. [ABSTRACT FROM AUTHOR]

A novel position‐aided fast millimetre‐wave beam training using compressive sensing (PAF‐CS) is proposed. In particular, the position information is used to obtain the angle‐of‐arrival, and this operation can facilitate beam alignment with significantly reduced system overhead. Aiming to resolve the position uncertainty, the PAF‐CS method first identifies a critical area by using the statistical information of the position error, and finds the best beams by using CS method. Theoretical analysis and simulation validate the superiority of the PAF‐CS method. [ABSTRACT FROM AUTHOR]

Quadrature spatial modulation (QSM) is one of the recently emerging transmission schemes to improve the throughput of the multi‐input multi‐output systems. QSM transmits an M‐ary complex symbol whose real and imaginary parts activate an antenna index independently thus provides an extra spatial dimension compared to conventional spatial modulation (SM). A generalised QSM (GQSM) transmission is considered in which a set of M‐ary complex symbols activate multiple antennas to transmit their real and imaginary parts separately, which offers higher throughput in terms of bits per channel use (bpcu) than other SM techniques. A detector is proposed, ordering‐based structured orthogonal matching pursuit for GQSM using compressive sensing (CS), which performs very close to the maximum likelihood detector with significantly lesser complexity. Also, the proposed algorithm performs superior to other CS‐based detectors and offers a better trade‐off between complexity and performance. [ABSTRACT FROM AUTHOR]

Previous research on orthogonal matching pursuit (OMP) algorithm mainly focuses on the recovery performance of a sparse signal x given an acquired model y=Fx+n. A general perturbation model y=(F+E)x+n in addition to the above acquired model exists, where E is the measurement perturbation. For this general perturbation model, the new restricted isometry constant of OMP algorithm is shown. [ABSTRACT FROM AUTHOR]

This Letter presents a sufficient condition under which the multipath matching pursuit (MMP) algorithm recovers sparse signals perfectly in compressive sensing problems. The derived sufficient condition is much more relaxed than the existing one. [ABSTRACT FROM AUTHOR]

A robust design method is proposed to optimise the measurement matrix for a compressive sensing multiple‐input multiple‐output (CS‐MIMO) radar system in the presence of clutter. The robustness to the proposed measurement matrix design is conducted against uncertainty on the target angle using a max–min optimisation problem to maximise the worst‐case signal‐to‐clutter‐plus‐noise ratio. The numerical results demonstrate the effectiveness of the proposed method in CS‐MIMO radar. [ABSTRACT FROM AUTHOR]

Practical acquisition of compressed sensing measurements involves a finite‐range finite‐precision quantisation step. To solve the sparse recovery problem and handle the quantisation distortion, this Letter proposes a non‐smooth graduated‐non‐convexity approach that follows a path of gradually improved solutions along a sequence of non‐smooth non‐convex optimisation problems that progressively promote quantisation consistency (QC) and sparsity. We consider two classes of multi‐scale continuous approximation functions to depict intermediate QC degrees and sparsity‐inducing strengths, respectively, and apply recent proximal splitting methods to solve the resulting subproblem at each refinement scale. The simulations demonstrate the convergence of intermediate solutions to a nearly optimal estimation, in terms of accuracy and support recovery. [ABSTRACT FROM AUTHOR]

Natural image reconstruction based on compressive sensing (CS) has shown a promising performance in recent years. However, sometimes the restoration precision is not high enough. A novel CS algorithm using vector quantisation (VQ) error is proposed. First, the original image is compressed by VQ due to its extremely high compression ratio and strong ability to preserve details. Then compute the VQ error matrix and ignore the three least significant bits, which makes the error matrix much sparser. Next, to ensure a uniform distribution of sparsity of blocks, the error matrix is scrambled. Since the huge diversity among blocks has been largely reduced, they can be sensed with the same sensing matrix in space domain. At last, the reconstruction effect of the error matrix decides the total restoration performance. Experimental results have demonstrated the proposed method, at low measurement ratio, performs better in the aspects of perception and peak signal‐to‐noise ratio. [ABSTRACT FROM AUTHOR]

Digital predistortion has become an attractive technique for power amplifier linearisation whose limiting factor for using Volterra series as the underlying model is its computational complexity, since the number of components rapidly grows with the non‐linear order and memory. Based on a previous reference algorithm, which consists on applying the orthogonal matching pursuit for the sorting of the model components and a Bayesian information criterion for the selection of the optimum number of components, a new technique to reduce the size of the support set taking into account the structural information within a model is presented. Experimental results of the predistortion of a commercial power amplifier are given as a proof of its capabilities, showing equivalent performance to the pruning with the reference algorithm while further reducing the number of components. [ABSTRACT FROM AUTHOR]

By block compressed sensing (BCS), the image can be compressed and encrypted simultaneously. A novel reversible data hiding (RDH) method is proposed to embed additional data into the BCS encrypted images. In the proposed method, based on the image block edge pixel separation, every block can be used for data embedding through a special measurement flipping. After decryption, considering the correlation of natural images, the fluctuation function can be used to find out the flipped block and realise the data extraction and image recovery. Experiments show that the proposed method can outperform the existing RDH method in BCS images. [ABSTRACT FROM AUTHOR]

The restricted isometry constant (RIC) bound for the recovery of a sparse signal with l2 bounded noise by orthogonal matching pursuit algorithm is studied. We show under our weaker condition on the RIC of sensing matrix and the minimum magnitude of the non‐zero components, the support of the unknown signal can be recovered exactly under l2 bounded noise. Our results are better than the best existing ones. [ABSTRACT FROM AUTHOR]

To overcome the disadvantages of the required motion compensation in the traditional inverse synthetic aperture radar imaging of moving targets, a 2D single snapshot imaging method using a MIMO radar based on the sparse representation (SR) is proposed. It can simultaneously recover a real‐time 2D image and avoid motion compensation. By exploiting the sparsity and data structure of the echo signals, a 2D SR model is established. The SR problem is solved based on the compressed sensing theory and the 2D smoothed L0 algorithm. Finally, the 2D image of the target can be reconstructed at low signal‐to‐noise ratio and high compressed ratio, which is advantageous for target recognition and classification. Compared with the existing 1D methods, the proposed method can directly recover 2D images and has lower computational complexity. Simulations are performed to verify the effectiveness of the method. [ABSTRACT FROM AUTHOR]

The support recovery performance of orthogonal matching pursuit (OMP) for the l∞ bounded noise is presented. It is shown that OMP can exactly recover the support of arbitrary K‐sparse signal x from noisy measurements y = Φx + e in K iterations, provided that the matrix Φ and the minimum magnitude of all the non‐zero elements of x satisfy certain requirements. The proposed result is an improvement over the existing ones. [ABSTRACT FROM AUTHOR]

A double sparse learning (DSL) model with a pyramid structure‐based feature extraction scheme to handle speech emotion recognition (SER) problem is proposed. The key novelty of the method is that the proposed DSL model is able to take into consideration two scales of the pyramid structure‐based features for selecting the features which have great contributions to SER. Extensive experiments on eNTERFACE and AFEW emotion databases to evaluate the method are conducted. The experimental results demonstrate that, compared with some recent competitive methods, DSL with the pyramid structure‐based feature extraction scheme has a more promising performance in dealing with the SER task. [ABSTRACT FROM AUTHOR]

A fast super‐resolution method for joint Doppler frequency, direction of departure and direction of arrival estimation in bistatic MIMO radar is presented. First, the 3D echo signal model of bistatic MIMO radar with uniform linear transmit and uniform linear receive array is established. Then, the Kronecker compressive sensing (KCS) is introduced to improve the 3D parameters estimation performance. Moreover, a novel fast orthogonal matching pursuit (OMP) algorithm utilising 3D fast Fourier transformation is also presented to reduce the computation burden of KCS. Finally, the effectiveness of the method is validated by the simulation. [ABSTRACT FROM AUTHOR]

A new methodology for interference/jamming suppression employing a spatial filter measurement matrix (SFMM) is presented. Different from the previous Capon beamforming method, the proposed method is not dependent on accurate prior information of targets' directions‐of‐arrival (DOAs). The proposed SFMM suppresses the noise and interferences outside the searching area, thus apparently improves detection performance. The experimental results demonstrate the effectiveness of the proposed SFMM. [ABSTRACT FROM AUTHOR]

A deterministic method to construct bipolar measurement matrices for compressed sensing is proposed based on Legendre sequences. The novel matrices have remarkably flexible measurement sizes, relatively low coherence and show empirically good performance compared with Gaussian matrices. [ABSTRACT FROM AUTHOR]