Search

To address the issue of performance degradation in the Capon beamformer when the desired signal appears in the training data, the integration operation is introduced for the interference-plus-noise covariance matrix reconstruction, and the steering vector (SV) is estimated by solving a convex optimization problem in some robust adaptive beamforming papers. However, this approach suffers from high computational complexity and is limited by the resolution of the Capon spectrum. In light of this, our paper proposes a novel robust adaptive beamforming method. To overcome the limited resolution of the Capon spectrum, we introduce the multiple signal classification method to acquire the nominal SVs. Subsequently, the SVs are updated based on subspace orthogonality. The proposed method constructs orthogonal components from the nominal SVs to circumvent the convex problem and reduce computational complexity. Additionally, the interference powers are obtained using the Capon spectrum without the noise component via eigenvalue decomposition. This refinement improves the accuracy of the estimated interference powers. Simulation results demonstrate that the proposed method is effective and robust against common mismatch errors. [ABSTRACT FROM AUTHOR]

This paper investigates the analysis of critical path delay in various infinite impulse response (IIR) digital filter structure implementations. The critical path delay increases with the order of filter increasing in all conventional structures. So, it is needed to reduce the critical path delay for the faster realization of a digital filter in real-time processing. This paper proposed the rules for taking efficient cutset to retime the higher-order IIR filter structure. By applying this proposed retiming/pipelining techniques to EEG preprocessing filter, the critical path delay is 28 % , 23 % and 18 % more reduced as compared to the traditional retiming in direct form, parallel and lattice-ladder structure, respectively. These newly retimed structures are simulated in MATLAB Simulink and converted to fixed-point arithmetic for the synthesis of structures on Virtex-5 implementation. The results in different structures show that the allpass-based IIR structures has the lowest critical path delay and lowest computational complexity as compared to the conventional structures. Also, the proposed retimed structure has a smaller slice-delay product and support the largest maximum sampling frequency but requires more slice lookup tables (LUTs) than a conventional structure. [ABSTRACT FROM AUTHOR]

Frequency domain filtered-x least mean square algorithms can reduce the computational complexity of the time domain counterpart with long filters; however, they suffer from large block delay, additional quantization error due to large size transformations and implementation difficulties in existing DSP hardware. In this paper, a time–frequency domain flexible structure is proposed using the partitioned block frequency domain adaptive filtering technique, which has no signal path delay and is well suited for low-cost DSP implementation. The proposed structure divides the long filters into many equal partitions and carries out the control filter update in frequency domain while generating the control signal in both time and frequency domains, thereby eliminating the forward path delay completely while maintaining low computational complexity. The proposed structure has a potential benefit for controlling broadband noise, where the causality constraint is more important. The simulation results using the measured acoustic paths demonstrate that the proposed structure maintains similar control performance as that of the time domain algorithm but with much less computational complexity. Furthermore, the tracking performance of the proposed structure under different levels of measurement noise is investigated. [ABSTRACT FROM AUTHOR]

Wideband linear frequency modulation (LFM) signals are widely used in radar, sonar, mobile, and similar applications. One of the common problems of algorithms for estimating the direction of arrival (DOA) of LFM signals is that the number of snapshots must be large for good estimation. Therefore, they are not suitable for real-time and low-power applications. In this paper, we proposed a method of estimating DOA based on the sparse iterative covariance (SPICE) algorithm, which has features such as low computational complexity for the uniform linear array (ULA) using the Fourier transform (FFT). We first developed the dechirp process for linear frequency modulation signals using the Fourier transform. Then we modified the SPICE algorithm for linear arrays. Finally, we have obtained the calculation of the DOA estimation for a few snapshots with low computational complexity and high resolution. Compared to other methods, the simulation results of the proposed LSPICE algorithm show an increase in estimation accuracy, higher resolution, more acceptable accuracy in low SNRs, less error in high SNRs, favorable response with low snapshots, and lower computational complexity. [ABSTRACT FROM AUTHOR]

In this paper, we demonstrate the significance of restoring harmonics of the fundamental frequency (pitch) in the deep neural network (DNN)-based speech enhancement. The parameters of the DNN can be estimated by minimizing the mask loss, but it does not restore the pitch harmonics, especially at higher frequencies. In this paper, we propose to restore the pitch harmonics in the spectral domain by minimizing cepstral loss around the pitch peak. Restoring the cepstral pitch peak, in turn, helps in restoring the pitch harmonics in the enhanced spectrum. The proposed cepstral pitch-peak loss acts as an adaptive comb filter on voiced segments and emphasizes the pitch harmonics in the speech spectrum. The network parameters are estimated using a combination of mask loss and cepstral pitch-peak loss. We show that this combination offers the complementary advantages of enhancing both the voiced and unvoiced regions. The DNN-based methods primarily rely on the network architecture, and hence, the prediction accuracy improves with the increasing complexity of the architecture. The lower complex models are essential for real-time processing systems. In this work, we propose a compact model using a sliding-window attention network (SWAN). The SWAN is trained to regress the spectral magnitude mask (SMM) from the noisy speech signal. Our experimental results demonstrate that the proposed approach achieves comparable performance with the state-of-the-art noncausal and causal speech enhancement methods with much lesser computational complexity. Our three-layered noncausal SWAN achieves 2.99 PESQ on the Valentini database with only 10 9 floating-point operations (FLOPs). [ABSTRACT FROM AUTHOR]

Model order reduction technique provides an effective way to reduce computational complexity in large-scale circuit simulations. This paper proposes a new model order reduction method for delay circuit systems based on Hermite expansion technique. The presented method consists of three steps i.e., first the delay elements are approximated using the recursive relation of Hermite polynomials, then in the second step, the reduced order is estimated for the delay circuit system using a delay truncation in the Hermite domain and in the third step, a multi-order Arnoldi process is computed for obtaining the projection matrix. In the following, the reduced order delay circuit model is obtained by the projection matrix. Moment matching and passivity properties of the reduced circuit system are also analyzed. Two circuit examples with delay components are performed to verify the effectiveness of the proposed MOR approach. [ABSTRACT FROM AUTHOR]

In this paper, a family of novel diffusion adaptive estimation algorithms is proposed from the asymmetric cost function perspective by combining diffusion strategy and the linear–linear cost, quadratic-quadratic cost, and linear-exponential cost at all distributed network nodes, and named diffusion LLCLMS (DLLCLMS), diffusion QQCLMS (DQQCLMS), and diffusion LECLMS (DLECLMS), respectively. Then, the stability of mean estimation error and computational complexity of those three diffusion algorithms are analyzed theoretically. Finally, several experiment simulation results are designed to verify the superiority of those three proposed diffusion algorithms. Results show that DLLCLMS, DQQCLMS, and DLECLMS algorithms are more robust to the input signal and impulsive noise than the diffusion sign-error LMS, diffusion robust variable step-size least mean square (DRVSSLMS), and least mean logarithmic absolute difference algorithms. In brief, theoretical analysis and experiment results show that those proposed DLLCLMS, DQQCLMS, and DLECLMS algorithms perform better when estimating the unknown linear system under the changeable impulsive noise environments and different environments types of input signals. [ABSTRACT FROM AUTHOR]

In this paper, the problem of single-channel blind source separation (SCBSS) is addressed using a novel approach that combines the adaptive mode separation-based wavelet transform (AMSWT) and the density-based clustering with sparse reconstruction. The proposed method is performed in the time–frequency domain and in a reverberant environment. First, using the Fourier transform, the amplitude spectrum of the observed mixture signal is obtained. Then, using variational scaling and wavelet functions, the AMSWT is used to adaptively extract spectral intrinsic components (SICs). To obtain a better time–frequency resolution, the AMSWT is applied to each mode. Thus, the SCBSS problem is transformed into a non-underdetermined. Then, for each frequency bin, the density-based clustering, reformulated to the eigenvector clustering problem, is performed to estimate the mixing matrix. Finally, sparse reconstruction is used to reconstruct the sources. The proposed method is evaluated using objective measures of separation quality. A computational complexity evaluation based on time consumption is also performed. Simulation results show that the proposed method is very effective for solving the SCBSS problem and provides better separation performances than the reference methods. However, the proposed method is computationally expensive. [ABSTRACT FROM AUTHOR]

This paper provides a detailed overview of active impulsive noise control (AINC) technology of α-stable distribution noise. The second-order moment of the non-Gaussian impulsive noise does not exist, so the adaptive algorithms based on second-order statistics would lead to poor noise reduction performance. On account of this defect, a series of AINC algorithms have been proposed over the past 25 years. In this paper, we briefly classified the AINC algorithms into the classic algorithms and the expanded algorithms depending on different criteria, and both the classic AINC algorithms and the expanded AINC algorithms are further divided. Afterward, we introduced the principles and characteristics of different AINC algorithms and listed the weights update formulas of those algorithms. Finally, the computational complexity comparisons of common AINC algorithms and further recommendations are given, and a brief review about the experimental testing of AINC systems is provided in order to assist related scholars to carry out AINC researches more conveniently. [ABSTRACT FROM AUTHOR]

Frequency estimation of sinusoid is an important research topic in the field of statistical signal processing. Efficient frequency estimation is of great significance for applications with high real-time requirements. In this paper, two efficient frequency estimators of complex sinusoid by using autocorrelation and discrete time Fourier transform (DTFT) are presented in additive white noise background. Firstly, autocorrelation algorithm is adopted to obtain the coarse frequency estimate. Then, three DTFT samples with arbitrary intervals located near the coarse estimate are used to get the fine frequency estimate. In order to evaluate the calculation complexity and estimation accuracy, the proposed estimators are compared with the competing algorithms. The computational complexity of the presented estimators is much lower than that of the competitive estimators. Simulation results show that the mean square errors of the presented estimators are close to the Cramer–Rao lower bound when the signal-to-noise ratio varies in a large range. [ABSTRACT FROM AUTHOR]

Based on the criterion of minimum error entropy, this paper proposes a novel conjugate gradient algorithm, called MEE-CG. This algorithm has robust performance under non-Gaussian interference. Theoretical analysis and experimental results demonstrate that the proposed algorithm displays more robust performance than the conventional conjugate gradient methods on the basis of the mean square error and the maximum correntropy criterion. Compared with the stochastic gradient minimum error entropy algorithm and the recursive minimum error entropy algorithm, the proposed algorithm provides a trade-off between computational complexity and convergence speed. [ABSTRACT FROM AUTHOR]

This paper presents an implementation of a reconfigurable digital down converter (DDC) that can translate high sample rate to lower sample rate signal on field-programmable gate array (FPGA) platform using polyphase filtering. The proposed DDC consists of a polyphase mixer, a cascaded integrator comb (CIC) filter, and a finite impulse response (FIR) filter. The polyphase mixer reduces computational complexity and multiplier blocks simultaneously. Furthermore, the modified CIC filter improves the operating speed, and the new design method of the FIR filter saves memory storage to a great extent. The proposed structure is highly flexible so that the decimation factor can be programmed dynamically during runtime in terms of the sample rate, bandwidth, center frequency, and so on. Moreover, the optimum hardware description language (HDL) coding techniques significantly improve the area efficiency and speed performance of the DDC without compromising the functionality. The proposed design is synthesized using Xilinx Vivado 2020.2 and tested on a Xilinx Kintex-7 FPGA XC7K70T-FBG676 as the target device. Comparison results indicate that the proposed design substantially reduces the resources as well as power. A verification test is to certify the feasibility and correctness of the hardware implementation. The proposed DDC is well-matched in any digital radio application. [ABSTRACT FROM AUTHOR]

The irrational coefficients of orthogonal wavelet filters require a large amount of hardware for their implementation. This results in huge memory requirement, increased power dissipation and reduction in speed of operation. This paper presents a new class of multiplier free db3 orthogonal wavelet filter bank with significantly reduced power dissipation, adders and shifters. In this paper, dyadic filter coefficients of db3 have been obtained by slightly altering double-shifting orthogonality property for perfect reconstruction. The proposed db3 wavelet filter improved time–frequency product, Sobolev regularity and frequency selectivity. Also, the VLSI architecture is suggested for the proposed db3 wavelet filter banks to analyse the hardware computational complexity. The performance of the proposed db3 wavelets achieved satisfactory performance for feature extraction in medical image retrieval and fingerprint recognition of infants and toddlers. [ABSTRACT FROM AUTHOR]

In this paper, we propose an $$l_0$$ -norm penalized shrinkage linear affine projection ( $$l_0$$ -SL-AP) algorithm and an $$l_0$$ -norm penalized shrinkage widely linear affine projection ( $$l_0$$ -SWL-AP) algorithm. The proposed algorithms provide variable step-size by minimizing the noise-free a posteriori error at each iteration and introduce an $$l_0$$ -norm constraint to the cost function. The $$l_0$$ -SWL-AP algorithm also exploits noncircular properties of the input signal. In contrast with conventional AP algorithms, the proposed algorithms increase the estimation accuracy for time-varying sparse system identification. A quantitative analysis of the convergence behavior for the $$l_0$$ -SWL-AP algorithm verifies the capabilities of the proposed algorithms. To reduce the complexity, we also introduce dichotomous coordinate descent (DCD) iterations to the proposed algorithms ( $$l_0$$ -SL-DCD-AP and $$l_0$$ -SWL-DCD-AP) in this paper. Simulations indicate that the $$l_0$$ -SL-AP and $$l_0$$ -SWL-AP algorithms provide faster convergence speed and lower steady-state misalignment than the previous APA-type algorithms. The $$l_0$$ -SL-DCD-AP and $$l_0$$ -SWL-DCD-AP algorithms perform similarly to their counterparts but with reduced complexity. [ABSTRACT FROM AUTHOR]

Motion estimation (ME) plays a vital role in performance of any video codec as it addresses the temporal redundancies existing within consecutive frames in any video sequence. The computational complexity of the ME unit is major with respect to the other blocks in the video codec. Researchers and practitioners prefer to use block matching algorithm for performing ME as it is efficient and simultaneously easy to implement. Some ME algorithms follow fixed search patterns, while the others are adaptive and tries to explore the correlation existing within neighbouring blocks of a video frame. This paper presents a new motion estimation algorithm named half-way stop adaptive pattern search algorithm (HAPSA) that reduces the computation of ME by lowering search points, using a proposed half-way stop technique and adaptive search pattern that can explore both fast and slow motion content in a video frame. The proposed HAPSA applies dual-threshold technique—one for classifying the block between zero motion and nonzero motion vector block and another for checking the accuracy of the motion vector. This helps in optimizing the clock cycle requirement and hardware utilization during VLSI design. The proposed VLSI system demonstrates architectural innovation that supports multiple dataflow models with high data throughput. The principal driving force behind the hardware architecture is an output stationary dataflow model wherein 16 reference frame pixels are broadcasted to all the processing elements every clock cycle. Working at a frequency of 357 MHz, the proposed architecture can process a 422 full HD (1920 × 1080) frames per second. The overall gate count of the design turns to be 49 K gate equivalent. The appreciable results with respect to some of the recent works can motivate one to include the proposed system in portable consumer video devices. [ABSTRACT FROM AUTHOR]

The generalized variable step-size diffusion continuous mixed p-norm (GVSS-DCMPN) algorithm is proposed in this paper, which is derived based on the improved continuous mixed p-norm (CMPN) strategy. In detail, a linear function is designed for the CMPN strategy. The proposed GVSS-DCMPN algorithm capable of exploiting various error norms to obtain performance improvement in non-Gaussian noise environment can be viewed as the generalization of the traditional p-norm-based algorithms in the sense of continuous errors. In particular, the GVSS-DCMPN algorithm transforms into the VSS-DCMPN algorithm when the slope of the linear function is set to 0. The computational complexity as well as the mean convergence is analyzed in the paper. Simulation results over the diffusion network show that the proposed algorithm achieves performance gain over some existing diffusion algorithms. [ABSTRACT FROM AUTHOR]

Graph filters are crucial tools in processing the spectrum of graph signals. In this paper, we propose to design universal IIR graph filters with low computational complexity by using three kinds of functions, which are Butterworth, Chebyshev, and elliptic functions, respectively. Specifically, inspired by the classical analog filter design method, we first derive the zeros and poles of graph frequency responses. With these zeros and poles, we construct the conjugate graph filters to design the Butterworth high-pass graph filter, Chebyshev high-pass graph filter, and elliptic high-pass graph filter, respectively. On this basis, we further propose to construct a desired graph filter of low pass, band pass, and band stop by mapping the parameters of the desired graph filter to those of the equivalent high-pass graph filter. Furthermore, we propose to set the graph filter order given the maximum passband attenuation and the minimum stopband attenuation. Our numerical results show that the proposed graph filter design methods realize the desired frequency response more accurately than the autoregressive moving average graph filter design method, the linear least-squares fitting-based graph filter design method, and the Chebyshev FIR polynomial graph filter design method. [ABSTRACT FROM AUTHOR]

Recently, compressed sensing algorithms, including convex optimization and greedy algorithm, are considered as a new development for direction of arrival (DOA) estimation. For the problem of two-dimensional (2-D) DOA estimation, the existing convex optimization-based methods are usually limited by their computational complexity, while the greedy algorithm-based estimation is subjected to its low estimation accuracy. In the current paper, we propose a new 2-D DOA estimation scheme with parallel coprime array. We proposed a modified orthogonal matching pursuit method for the parallel coprime array, which transforms the 2-D DOA estimation into one-dimensional (1-D) DOA estimation to reduce the computation complexity. In addition, the proposed estimation scheme makes full use of the cross-correlation information of the received signal and combines the advantages of the coprime array to achieve a much higher precision with fewer array elements. The simulation results demonstrate the superiority of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Complex harmonic wavelet (CHW) is a time–frequency tool that directly computes the frequency content rather than scales with respect to time. The complex harmonic wavelet for discrete signal computes N coefficients for N input samples, with the computational complexity of O (N log 2 N) by employing the FFT algorithm. However, this wavelet function for discrete signal is not conjugate symmetric for positive and negative half-planes. Therefore, this paper proposes a new basis function termed modified complex harmonic wavelet transform by multiplying a phase term with CHW. The novelty of this phase multiplication leads the modified CHW to complex conjugate that reduces its computational complexity compared to existing CHW, attaining a computational gain up to 25 % asymptotically for larger sample points. Like CHW, this proposed modified CHW retains the properties of orthogonality and compact support in the frequency domain. [ABSTRACT FROM AUTHOR]

In this paper, a novel diffusion estimation algorithm is proposed from a probabilistic perspective by combining the diffusion strategy and the probabilistic least mean square (LMS) at all distributed network nodes. The proposed method, namely diffusion-probabilistic LMS (DPLMS), is more robust to the input signal and impulsive noise than previous algorithms like the diffusion sign-error LMS (DSE-LMS), diffusion robust variable step-size LMS (DRVSSLMS), and diffusion least logarithmic absolute difference (DLLAD) algorithms. Instead of minimizing the estimation error, the DPLMS algorithm is based on approximating the posterior distribution with an isotropic Gaussian distribution. In this paper, the stability of the mean estimation error and the computational complexity of the DPLMS algorithm are analyzed theoretically. Simulation experiments are conducted to explore the mean estimation error for the DPLMS algorithm with varied conditions for input signals and impulsive interferences, compared to the DSE-LMS, DRVSSLMS, and DLLAD algorithms. Both results from the theoretical analysis and simulation suggest that the DPLMS algorithm has superior performance than the DSE-LMS, DRVSSLMS, and DLLAD algorithms when estimating the unknown linear system under the changeable impulsive noise environments. [ABSTRACT FROM AUTHOR]

As technology shrinks, multiple bits upsets (MBUs) are becoming an important problem in the reliability of digital designs exposed to radiation effects. In this paper, a new technique for the implementation of finite impulse response (FIR) filters is presented that provides protection against single and multi-unidirectional bit upsets (SEUs and MBUs), which have a lower circuit complexity and cost than traditional techniques like N-modular redundancy. Most of previous works has focused on single event upset (SEU), however, in this paper, a coding method based on Berger codes are presented and implemented on FPGA-based FIR filter. As the Berger coding covers all unidirectional faults, the MBUs are also handled in the proposed schemes. The effectiveness of the proposed technique has been evaluated using a dynamic partial reconfiguration-based fault injection platform. The implementation results of the proposed mitigation technique in comparison with traditional TMR method and previous mitigation techniques in the literature represent the effectiveness in terms of protection and implementation cost. [ABSTRACT FROM AUTHOR]

In this paper, we study the problem of estimation of parameters of multichannel sinusoidal model. In multichannel sinusoidal model, the inherent frequencies from distinct channels are same with different amplitudes. It is assumed that the errors in individual channel are independently and identically distributed, whereas the signal from different channels is correlated. We first propose to minimize the sum of residual sum of squares to estimate the unknown parameters, and they can be easily obtained. Next we propose to use more efficient generalized least squares estimators which become the maximum likelihood estimators also when the errors follow multivariate Gaussian distribution. Both the estimators are strongly consistent and asymptotically normally distributed. We have provided the implementation of the generalized least squares estimators. Simulation experiments have been performed to compare the performances of the least squares estimators and generalized least squares estimators. It is observed that the variances of the maximum likelihood estimators reach the Cramer–Rao lower bound even for moderate sample sizes. We have extended the methods of estimation and the associated results of the two-channel model to an arbitrary m-channel model. It is observed that the computational complexity does not increase significantly with the increase in number of channels. [ABSTRACT FROM AUTHOR]

This paper extends the concept of Karnaugh map from a logical function to a logical system, thereby allowing both the state transition graph and state transition matrix of a logical system, as well as a digital circuit, to be constructed directly. It is found that the Karnaugh map of a logical system is actually a two-dimensional state transition table or, equivalently, a table-like state transition graph and is also a grid-type state transition matrix. For constructing the state transition graph and state transition matrix of a logical system, the computational complexities of the methods based on the proposed Karnaugh map are exponentially lower than those of the existing methods. The Karnaugh maps of logical systems are then used in the analysis and design of clocked sequential circuits and in the simplification of multioutput gate circuits. Some illustrative examples and simulations are presented to demonstrate the results and their applications. [ABSTRACT FROM AUTHOR]

Polynomial matrix computations, such as polynomial matrix multiplication (PMM) and eigenvalue factorization of parahermitian matrices, have played an important role in a growing number of applications, in recent times. However, the computational complexity and expense of such operations impose a profound limit on their applicability. In a recent paper, we introduced a systolic array-based parallel architecture for PMM, which was adequately efficient, but limited in its application. In this paper, we propose a second-generation hardware solution which boasts more versatility, efficiency and scalability compared to our previous design. This is achieved through the design of a highly versatile PMM accelerator which supports polynomial matrices of any size, as a component of the embedded system developed within the Xilinx Zynq-7000 AP SoC. Experimental results demonstrate the efficiency and effectiveness of our novel SoC-based PMM accelerator in the context of subband coding, where maximum speedups of 85 × and 33 × are accomplished, without compromising the accuracy, in comparison with two highly optimized and multi-threaded software-only implementations running on a dual-core ARM Cortex-A9 processor and a Intel Core i7-4510U CPU, respectively. [ABSTRACT FROM AUTHOR]

Adaptive filters based on least-mean-square (LMS) algorithm are used in several applications in virtue of their good steady-state performance, numerical stability, and acceptable computational complexity. The hardware implementation of LMS filters requires a massive number of multipliers that significantly impact on the power consumption. Approximate computing, a design technique that trades off computation accuracy for better electrical performance, is a way to improve the energy efficiency of LMS filters. In this paper, we implement state-of-the-art approximate multipliers and evaluate their impact on the performance of the LMS algorithm. Moreover, a novel approximate multiplier, whose accuracy can be tuned at design time to better adapt to the application scenario, is proposed. Implementation results in 28-nm CMOS technology allow us to investigate the power versus quality trade-off of the considered LMS approximate filters in two different applications. [ABSTRACT FROM AUTHOR]

The complex harmonic wavelet (CHW) for the discrete signal is orthogonal, compact support in the frequency domain but is not complex conjugate symmetric for positive and negative half-planes. The modified complex harmonic wavelet (MCHW) transform is the improved version of CHW as it is the complex conjugate symmetry along with other desired properties of CHW such as orthogonality, and compact support in the frequency domain. Due to the complex conjugate symmetry, MCHW has lesser computational complexity compared to CHW. This paper introduces a new VLSI architecture for MCHW for hardware implementation and prototyped on a commercially available virtex5 field-programmable gate array (FPGA). For the validation of the proposed implementation, the real-time captured results in the logic analyzer are verified with simulation results. The maximum operating frequency targeting the above-mentioned FPGA device is reported as 92.82 MHz. The total on-chip power of the above implementation is 1.117W, out of which 84 mW is the dynamic power dissipation at a toggle rate of 12.5 %. Finally, for the area utilization of the above implementation, its resource utilization targeting the above FPGA device is reported. [ABSTRACT FROM AUTHOR]

Recent research on sparse signal reconstruction (SSR) shows that the sparse Bayesian learning (SBL) method is one of the most popular methods for direction-of-arrival (DOA) estimation. However, the high computational complexity and the time-consuming iterations in the Bayesian learning make the SBL-based methods difficult in practical applications. To address this problem, the iterative process in the SBL method is improved in this paper. Inspired by the spatial sparsity of the signal variance, an l 0 norm constraint Bayesian strategy for DOA estimation is proposed. An additional l 0 norm penalty is integrated into the objective function of the signal variance vector to achieve zero attraction. With a few user-adjusted parameters, the proposed strategy achieves a faster implementation while inheriting the superior performance of the SBL method for DOA estimation. Numerical simulations demonstrate that the proposed strategy can significantly improve the convergence rate and accelerate the iterative process. The innovative work increases the possibility of real-time implementation. [ABSTRACT FROM AUTHOR]

In this article, a widely linear complex-valued affine projection sign algorithm is proposed for handling nonstationary noise in the complex domain. Apart from using the second-order statistical information in the complex domain adequately, the proposed algorithm can succeed in obtaining the robustness against the impulsive noise. With the help of the Price's theorem and some rational assumptions, the steady-state mean-square analysis of the proposed algorithm has been shown in this paper. In addition, we achieve a low computational complexity combining the l1-norm operation of the error signal. Moreover, we still analyze the stability of our finding, providing the range of the step size of its widely linear model. In the end, the results of experiments with the impulsive noise make clear that the proposed algorithm obtains the robustness for both complex circular and noncircular signals. However, simulation results without the impulsive noise indicate that the proposed algorithm gets a worse behavior in terms of normalized mean-square deviation compared with various existing algorithms. In the stereophonic acoustic echo cancellation, our finding still achieves a good performance. [ABSTRACT FROM AUTHOR]

The cubature Kalman filter (CKF) and its square-root version, namely the square-root cubature Kalman filter (SCKF), are commonly used nonlinear estimators under Gaussian noises. Particularly, the latter has the added advantage of low computational complexity and guaranteed positive semi-definiteness. However, the estimation accuracy often degrades substantially when the measurements are contaminated by outliers. This paper proposes a robust Laplace ℓ 1 square-root cubature Kalman filter (LSCKF). The proposed filter employs the heavy-tailed Laplace distribution to model the measurement noises and solves the maximum posterior estimation using the majorization minimization (MM) approach and Gauss–Newton method. Besides, the filter is derived in square root and uses the orthogonal transformations to realize reliable computation of state estimates. Therefore, the new filter not only has good robustness against measurement outliers, but also retains the advantages of high numerical stability. Two numerical simulations are performed to test the performance of the proposed algorithm. [ABSTRACT FROM AUTHOR]

The reconstruction of time-varying signals on graphs is a prominent problem in graph signal processing community. By imposing the smoothness regularization over the time-vertex domain, the reconstruction problem can be formulated into an unconstrained optimization problem that minimizes the weighted sum of the data fidelity term and regularization term. In this paper, we propose an approximate Newton method to solve the problem in a distributed manner, which is applicable for spatially distributed systems consisting of agents with limited computation and communication capacity. The algorithm has low computational complexity while nearly maintains the fast convergence of the second-order methods, which is evidently better than the existing reconstruction algorithm based on the gradient descent method. The convergence of the proposed algorithm is explicitly proved. Numerical results verify the validity and fast convergence of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Speech is a user-friendly signal for identity recognition with low computational complexity and implementation cost. However, the use of speech samples to identify persons involves several limitations, such as degraded performance in real environments due to the presence of different noises and channel effects. In recent years, deep neural network (DNN)-based approaches have provided good results in speaker verification and outperformed the i-vector based methods. The x-vector is a DNN-based speaker embedding that, in combination with probabilistic linear discriminant analysis (PLDA), increases both the accuracy and robustness of speaker verification systems. In this paper, we propose weighted x-vectors as a method for enhancing the speaker verification system in both clean and noisy environments. It exploits the statistical properties of target speaker enrollment x-vectors for weighting the test x-vector to enhance the scoring accuracy and thus the whole verification system. Experiments were conducted using the VoxCeleb dataset, MFCC feature vectors, and PLDA scoring method. The VoxCeleb is a large-scale dataset that contains real-world short-duration speech samples from over 6,000 speakers. Multicondition training for LDA and PLDA was also employed to improve the system's performance under mismatched noisy circumstances. The findings showed that using weighted x-vectors led to 18% and 10% reductions in equal error rate (EER) term for clean and noisy conditions, respectively. Also, the experiments show that the increase of the number of enrollment x-vectors results in superior performance of the proposed method. [ABSTRACT FROM AUTHOR]

In the direction-of-arrival (DOA) estimation with a coprime linear array, the ambiguity problems such as pair matching and grating angle have been investigated. In the ambiguity problem situation, the estimation reliability and accuracy of the existing subarray and adjoined-subarray-based methods degrade due to the low angular resolution. Moreover, it entails excessive computational complexity and execution time. In this paper, to tackle these limitations, DOA estimation is handled as approximating the unknown regression function that maps the eigenvector of signal subspace with the DOA of the incoming source signals using propagator support vector regression. The proposed method offers good generalization and robustness in resolving the ambiguity problem. It achieves full degrees of freedom (DOF) with reduced complexity. To demonstrate the superiority and effectiveness of the proposed method, various simulations in terms of estimation reliability, accuracy, DOF and complexity analysis are provided in comparison with the existing methods. [ABSTRACT FROM AUTHOR]

Detection and recovery of noise corrupted signals using adaptive filters are becoming popular due to their application in several fields, including communication and biomedical. Adaptive filters broadly utilize the LMS algorithm since it has low computational complexity and is robust in behavior. However, it fails to achieve a faster speed of convergence and minimum steady-state MSE simultaneously. Therefore, the objective is to estimate the noise-free signal faster with improved accuracy and obtain a lower steady-state MSE with a higher convergence speed. This paper introduces a multi-stage adaptive filtering model wherein the noisy signal is analyzed through a cascade of stages. The proposed signal de-noising scheme utilizes the automatic selection of stages to be cascaded such that the steady-state MSE is minimum. Further, to enhance convergence speed, the LMS adaptive filter's step-size is automatically adjusted at each stage. Hence, by controlling the number of stages to be cascaded automatically and utilizing a different step-size for each stage, we obtain a high convergence speed and minimum steady-state MSE. The proposed filter structure is tested for signal de-noising application to assess its performance concerning MSE, Signal-to-Noise Ratio (SNR), Average Noise Reduction (ANR), and convergence speed. The results attained have shown that the proposed filter structure provides remarkable performance improvement at different input noise levels. Further, the proposed Improved Variable-Stage (IVS) cascaded adaptive filter model employs LMS adaptive algorithm, hence offering a cost-effective and straightforward hardware implementation of ANC. [ABSTRACT FROM AUTHOR]

With low hardware cost and power consumption, direction-of-arrival (DOA) estimation exploiting one-bit quantized array data has become an attractive topic. In this paper, the gridless compressed sensing (CS) with multiple measurement vectors for one-bit DOA estimation is investigated. First, the one-bit quantized signal model is established. Then an atomic norm minimization scheme is proposed based on the output of one-bit quantizer, in which the objective function is reformulated as a semidefinite programming problem combined with a one-sided l 1 -norm constraint, representing the sign inconsistency between the quantized and unquantized measurements. To efficiently solve such a problem and reduce its computational complexity, an accelerated proximal gradient-based algorithm is developed. The proposed approach outperforms the one-bit MUSIC in a small number of measurements, and avoids the grid-mismatch issue of several existing one-bit CS methods. Numerical experiments are conducted to validate the superiorities of the proposed one-bit DOA estimation approach in accuracy and running time. [ABSTRACT FROM AUTHOR]

Compressed sensing (CS)-based cross-and-bouquet (CAB) model was proposed by J. Wright et al. to reduce the complexity of sparse error correcting. For the sake of leading to better performance of CS-based decoding for the CAB model, an algorithm is proposed in this paper for constructing a well-designed projection matrix to minimize the average measures of mutual coherence. One was proposed by M. Elad. Another is defined in this paper for higher dimensional cases. Using the equivalent dictionary, the dimensionality is reduced. Also high-dimensional singular value decomposition (SVD) is avoided in the procedure of constructing a well-designed projection matrix. The high-dimensional CAB model of sparse error correcting can be solved by the proposed algorithm without computational difficulty. The validity of the proposed method is illustrated by decoding experiments in high-dimensional cases. [ABSTRACT FROM AUTHOR]

This paper deals with reduction of computational complexities in dynamic systems. This paper develops a novel method of reducing complexities with use of control moments of the system. Though the proposed method is validated through channel estimation in this paper, the same can be equally applied to any other dynamic systems. Encouraging results given in this paper prove that the computational complexities can be reduced up to 10 with a marginal affordable loss of performance. [ABSTRACT FROM AUTHOR]

The sparse signal recovery-based direction of arrival (DOA) estimation has received a great deal of attention over the past decade. From the sparse representation point of view, ℓ0-norm is the best choice to evaluate the sparsity of a vector. However, solving an ℓ0-norm minimization problem is non-deterministic polynomial hard (NP-hard). Thus, The common idea for many sparse DOA estimation methods is to use the ℓ1-norm as the sparsity metric. However, its sparse solution may not coincide with the solution resulting from the ℓ0-norm thus deteriorating the DOA estimation performance. In this paper, we propose a new sparse method based on ℓp (0

The purpose of kernel adaptive filtering (KAF) is to map input samples into reproducing kernel Hilbert spaces and use the stochastic gradient approximation to address learning problems. However, the growth of the weighted networks for KAF based on existing kernel functions leads to high computational complexity. This paper introduces a reduced Gaussian kernel that is a finite-order Taylor expansion of a decomposed Gaussian kernel. The corresponding reduced Gaussian kernel least-mean-square (RGKLMS) algorithm is derived. The proposed algorithm avoids the sustained growth of the weighted network in a nonstationary environment via an implicit feature map. To verify the performance of the proposed algorithm, extensive simulations are conducted based on scenarios involving time-series prediction and nonlinear channel equalization, thereby proving that the RGKLMS algorithm is a universal approximator under suitable conditions. The simulation results also demonstrate that the RGKLMS algorithm can exhibit a comparable steady-state mean-square-error performance with a much lower computational complexity compared with other algorithms. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel low-complexity scheme to extract cyclic features of wideband signals from sub-Nyquist samples as Nyquist rates push contemporary analog-to-digital converters to their performance limits. Due to the sparse spectrum occupancy in wideband, sub-Nyquist sampling is performed and cyclostationary feature extraction is achieved at baseband to identify the modulation scheme with low computational complexity. The multichannel sub-Nyquist sampling structure of the modulated wideband converter (MWC) is used to perform spectrum sensing. Automatic modulation identification and classification are done based on cyclostationary feature detection for the efficient use of available spectrum bands. The computational complexity is reduced as the aliased copy of the signal in baseband obtained in each channel of MWC is only used to estimate the cyclic spectrum of the signal and identify the modulation scheme. Simulations performed validate the proposed method for various modulation schemes. [ABSTRACT FROM AUTHOR]

This paper aims to extend the proportionate adaptation concept to the design of a class of diffusion normalized subband adaptive filter (DNSAF) algorithms. This leads to four extensions of the algorithm associated with different step-size variations, namely diffusion proportionate normalized subband adaptive filter (DPNSAF), diffusion μ -law PNSAF (DMPNSAF), diffusion improved PNSAF (DIPNSAF) and diffusion improved IPNSAF (DIIPNSAF). Subsequently, steady-state performance, stability conditions and computational complexity of the proposed algorithms are investigated. For each extension the performance has been evaluated using both real and simulated data, where the outcomes demonstrate the accuracy of the theoretical expressions and effectiveness of the proposed algorithms. [ABSTRACT FROM AUTHOR]

In recent years, the convolutional neural networks (CNNs) have been applied to many fields due to its high performance for extracting complex features. However, these CNNs models are robust but come at the cost of lots of computational complexity. As a result, a bunch of studies researched the various architectures and data flows for optimizing the throughput and energy efficiency. This paper presents a reused data flow and a shift and difference-add booth multiplier to reduce energy consumption. The evaluation result uses the pre-trained VGG16 model with a batch size of three as a benchmark. The result shows that the proposed design reduces the number of state toggles in the booth multiplier by 1.96 times and reduces the DRAM and global buffer accesses to 61.6% and 74.7% as prior work, respectively. [ABSTRACT FROM AUTHOR]

In this paper, a two-channel critically sampled spline graph filter bank (SGFB) for an arbitrary undirected graph with spectral domain sampling is proposed. The filters in the analysis section satisfy perfect reconstruction condition, and the synthesis section is implemented with low computational complexity. The proposed SGFB maintains the spectrum of the original signal in multi-level decomposition and then suppresses noise efficiently. It is worth noting that the proposed filter design is performed once for the desired filter shape in the spectral domain with desired cutoff frequency. Since the optimization problem is solved once, it reduces the complexity for large graph significantly. Through numerical analyses, we validate the efficacy of the proposed SGFB using multi-level decomposition of a graph signal and noise suppression in terms of signal-to-noise ratio (SNR) improvement. [ABSTRACT FROM AUTHOR]

Sinusoidal model and chirp model are the two fundamental models in digital signal processing. Recently, a chirp-like model was introduced by Grover et al. (International conference on computing, power and communication technologies, IEEE, pp. 1095–1100, 2018). A chirp-like model is a generalization of a sinusoidal model and provides an alternative to a chirp model. We derive, in this paper, the asymptotic properties of least squares estimators and sequential least squares estimators of the parameters of a chirp-like signal model. It is observed theoretically as well as through extensive numerical computations that the sequential least squares estimators perform at par with the usual least squares estimators. The computational complexity involved in the sequential algorithm is significantly lower than that involved in calculating the least squares estimators. This is achieved by exploiting the orthogonality structure of the different components of the underlying model. The performances of both the estimators for finite sample sizes are illustrated by simulation results. In the specific real-life data analyses of signals, we show that a chirp-like signal model is capable of modeling phenomena that can be otherwise modeled by a chirp signal model, in a computationally more efficient manner. [ABSTRACT FROM AUTHOR]

This paper addresses the problem of two-dimensional (2D) direction of arrival estimation of coherent sources with less computational complexity using sparse arrays. The proposed method incorporates the unitary ESPRIT method into three parallel nested and coprime arrays configuration. The advantages of the proposed method are computationally efficient due to real-valued operations, automatic pairing of 2D angles, resolves coherent targets, large array aperture and less mutual coupling. Simulation results show the effectiveness of the proposed method. Cramer–Rao lower bound (CRLB) has been given for the reference. The performance of the proposed method is compared with the methods available in the existing literature. The results achieved by the proposed method were found to be significant in terms of estimation accuracy, resolution, and the computational complexity. [ABSTRACT FROM AUTHOR]

This paper focuses on the high resolution time-frequency distribution (TFD) of multicomponent signals with amplitude and frequency modulations, and a concise method named short-time sparse representation (STSR) is proposed. In STSR, both analysis and synthesis of the discrete signal can be achieved by exploiting the signal's sparsity in frequency domain at each time instant. In order to fasten the STSR procedure, an efficient sparse recovery algorithm named SL0 is applied, and the signal model for each sliding window is modified to form the same dictionary, which guarantees that the whole recovery procedure adapts to the matrix form. The performance of STSR is compared with other TFD techniques and assessed in various configurations. It is shown that both preferable representation and acceptable computational cost can be obtained. [ABSTRACT FROM AUTHOR]

The effectiveness of communication system is impaired by the carrier frequency offset (CFO) and inter-carrier interference (ICI) due to the significant time variation under doubly selective (DS) channels. In this paper, we consider the weighted-type fractional Fourier transform (WFRFT)-based hybrid carrier modulation (HCM) system with partial fast Fourier transform (PFFT) demodulation for DS channels in presence of CFO. The influence of CFO and ICI, after PFFT demodulation and optimal channel compensation, can be further suppressed under certain WFRFT domain. Numerical results are presented to demonstrate the superiority of HCM with PFFT demodulation over the single carrier modulation and orthogonal frequency division multiplexing systems with PFFT demodulation at a very moderate complexity. [ABSTRACT FROM AUTHOR]

In this paper, the performance loss caused by the aperture fill phenomena and heavy computational burden in broadband adaptive beamforming (ADBF) is analyzed, and an efficient approach based on frequency-space cascade processing is presented. First, a digital compensation method is applied to compensate for the aperture fill time of the broadband phased array in the range-frequency domain, removing the dependence of the wideband array steering vector on the instantaneous frequency. Second, a subband adaptive beamforming scheme using range-compressed data is implemented, reducing the degrees of freedom of wideband jamming and avoiding signal cancelation in the ADBF covariance matrix estimation. Moreover, the conjugate gradient algorithm is used to iteratively calculate the ADBF weight vector to reduce the computational complexity. Finally, wideband ADBF can be achieved by subband synthesis. Theoretical analysis and simulation results are provided to demonstrate that the proposed approach can suppress jamming sufficiently with a rapid convergence rate, making this approach feasible for engineering implementation. [ABSTRACT FROM AUTHOR]

This paper presents a new split-radix algorithm for DHT of length $${N}=2^{n}$$ , called the Dual Split-Radix DHT (DSR DHT), that allows an efficient parallel implementation using a dual core system. Moreover, as it is different from existing split-radix algorithms for DHT, it offers an efficient hardware implementation similar to that for FFT. It avoids the so-called retrograde indexing specific to existing DHT algorithms that do not allow an efficient pipeline implementation. [ABSTRACT FROM AUTHOR]

Digital video technology has been increasingly needed in various fields, such as telecommunications, entertainment, medicine. Therefore, video compression is required. Motion estimation methods help in improving video compression efficiency by effectively removing the temporal redundancy between successive frames. Several block-based motion estimation (BME) algorithms are being suggested to reduce the coding process's computational complexity. This paper proposes a new rapid hybrid (BME) algorithm established on the primary search point prediction and advance ending search point strategies. It combines rough adaptive search and effective local search. The coarse search introduces a new motion vector (MV) prediction technique that utilizes the macro-blocks (MBs) Spatio-temporal correlations to optimize the traditional adaptive-rood-pattern search algorithm (ARPS) and speeding up the whole process without affecting the accuracy. In the accurate local search, the cross-formed search pattern using a one-step search (OSS) block matching algorithm is employed, to estimate the actual (MV) with less computation time and further speed up the search efficiency. Exhaustive experiments are performed to demonstrate the present algorithm's performance over the benchmark schemes concerning specific assessment criteria for results, including the peak signal-to-noise ratio (PSNR), computational complexity and computational gain. The results show that the proposed algorithm is efficient and reliable; it can always give better performance over diamond search (DS) and (ARPS). The conducted test shows an increased performance of search speed while preserving the visual quality of the motion-compensated images, and it achieves 59.76–88.03 speed improvement over (DS) and 20.98–72.06 over (ARPS) for different video sequences. Besides, the suggested method (ARP-OSS) provides the best result compared to (DS) and (ARPS) in terms of time complexity for analyzing all the video samples. [ABSTRACT FROM AUTHOR]

This paper proposes a novel multicycle spectrum sensing method. The method first divides the received data samples into several segments, and then gets cyclic autocorrelation estimates from these segments, constructs a test statistic from these estimates, and finally makes an F-test to determine whether there is a primary signal or not. The method can use multiple cyclic frequencies simultaneously; hence, it is a multicycle detector. The method is also nonparametric, and thus it is robust to noise uncertainty. Compared with existing cyclic spectrum sensing methods, the proposed method can reduce computational complexity significantly at the cost of a little performance loss. Simulation results are given to show the validity and the superiority of the proposed method. [ABSTRACT FROM AUTHOR]