检索结果-内蒙古大学图书馆

Sky Imager Data Reduction Using Autoencoder and Internet of Things Computing

IEEE ACCESS 2022年 10卷 111232-111240页

作者： Alotaibi, Ziyad N. Khouli, Saeed A. Ibrahim, Ahmed B. Alrabeiah, Muhammad Ragheb, Amr M. Almaiman, Ahmed S. Alshebeili, Saleh A. King Saud Univ Elect Engn Dept Riyadh 11421 Saudi Arabia

Forecasting the output power of renewable energy sources is of an utmost importance for maintaining the stability of the electricity grid and ensuring uninterrupted service. Internet-of-Things (IoT) devices equipped with sky imagers are often used to estimate solar irradiance whose value is essential for forecasting solar energy. This paper considers distributed sky imagers whose data are transmitted through the Internet to a cloud for possible storage, forecasting, and/or performing other signal processing operations. The data comprise a collection of images, and this makes their streaming to the cloud a burden, i.e., they require a large bandwidth. In an attempt to deal with that, an IoT system is proposed in this paper such that it collects, compresses, transmits, and reconstructs those images with as little bandwidth as possible. More specifically, an IoT device multiplexes the collected images and compresses them using a deep learning algorithm, the encoder of a convolutional autoencoder (CAE) in particular. It, then, transmits the compressed images to a cloud server where they could be decompressed with high-fidelity using the decoder of the CAE. As the core to the proposed system, the performance of the CAE algorithm is evaluated using a publicly available dataset of sky images. Results indicate that the images can be compressed to 2% of their original size (i.e., compression ratio of 98%) and reconstructed again with high fidelity, i.e., average structural similarity index measure (SSIM) of 99%. Hardware implementation of the CAE algorithm has also been attempted using a Raspberry Pi platform representing the IoT device with an average processing speed of 6.6681 seconds per frame. This speed well meets practical requirements given the low frame-rate of sky imagers.

关键词： Image coding Forecasting Internet of Things Cloud computing signal processing algorithms Codecs Image reconstruction Convolutional neural networks Convolutional auto-encoder image compression sky imager raspberry pi Internet of Things

来源：评论

学校读者我要写书评

暂无评论

Over-Parametrized Matrix Factorization in the Presence of Spurious Stationary Points

引用

IEEE TRANSACTIONS ON signal processing 2022年 70卷 482-496页

作者： Eftekhari, Armin Umea Univ Dept Math & Math Stat S-90187 Umea Sweden

Motivated by the emerging role of interpolating machines in signal processing and machine learning, this work considers the computational aspects of over-parametrized matrix factorization. In this context, the optimization landscape may contain spurious stationary points (SSPs), which are proved to be full-rank matrices. The presence of these SSPs means that it is impossible to hope for any global guarantees in over-parametrized matrix factorization. For example, when initialized at an SSP, the gradient flow will be trapped there forever. Nevertheless, despite these SSPs, we establish in this work that the gradient flow of the corresponding merit function converges to a global minimizer, provided that its initialization is rank-deficient and sufficiently close to the feasible set of the optimization problem. We numerically observe that a heuristic discretization of the proposed gradient flow, inspired by primal-dual algorithms, is successful when initialized randomly. Our result is in sharp contrast with the local refinement methods which require an initialization close to the optimal set of the optimization problem. More specifically, we successfully avoid the traps set by the SSPs because the gradient flow remains rank-deficient at all times, and not because there are no SSPs nearby. The latter is the case for the local refinement methods. Moreover, the widely-used restricted isometry property plays no role in our main result.

关键词： Interpolation Optimization Neural networks signal processing algorithms signal processing Training Toy manufacturing industry Rank-constrained matrix factorization Burer-Monteiro factorization over-parametrization interpolation nonconvex optimization stationary points

来源：评论

学校读者我要写书评

暂无评论

DDPG-Driven Deep-Unfolding With Adaptive Depth for Channel Estimation With Sparse Bayesian Learning

引用

IEEE TRANSACTIONS ON signal processing 2022年 70卷 4665-4680页

作者： Hu, Qiyu Shi, Shuhan Cai, Yunlong Yu, Guanding Zhejiang Univ Coll Informat Sci & Elect Engn Hangzhou 310027 Peoples R China Zhejiang Univ Zhejiang Prov Key Lab Informat Proc Commun & Netw Hangzhou 310027 Peoples R China

Deep-unfolding neural networks (NNs) have received great attention since they achieve satisfactory performance with relatively low complexity. Typically, these deep-unfolding NNs are restricted to a fixed-depth for all inputs. However, the optimal number of layers required for convergence changes with different inputs. In this paper, we first develop a framework of deep deterministic policy gradient (DDPG)-driven deep-unfolding with adaptive depth for different inputs, where the trainable parameters of deep-unfolding NN are learned by DDPG, rather than updated by the stochastic gradient descent algorithm directly. Specifically, the optimization variables, trainable parameters, and architecture of deep-unfolding NN are designed as the state, action, and state transition of DDPG, respectively. Then, this framework is employed to deal with the channel estimation problem in massive multiple-input multiple-output systems. Specifically, first of all we formulate the channel estimation problem with an off-grid basis and develop a sparse Bayesian learning (SBL)-based algorithm to solve it. Secondly, the SBL-based algorithm is unfolded into a layer-wise structure with a set of introduced trainable parameters. Thirdly, the proposed DDPG-driven deep-unfolding framework is employed to solve this channel estimation problem based on the unfolded structure of the SBL-based algorithm. To realize adaptive depth, we design the halting score to indicate when to stop, which is a function of the channel reconstruction error. Furthermore, the proposed framework is extended to realize the adaptive depth of the general deep neural networks (DNNs). Simulation results show that the proposed algorithm outperforms the conventional optimization algorithms and DNNs with fixed depth with much reduced number of layers.

关键词： Channel estimation Artificial neural networks signal processing algorithms Adaptive systems Optimization Iterative methods Discrete Fourier transforms Deep deterministic policy gradient deep unfolding with adaptive depth channel estimation sparse Bayesian learning massive MIMO systems

来源：评论

学校读者我要写书评

暂无评论

signal processing with Spiking Neuron Model

Signal Processing with Spiking Neuron Model

引用

Digital signal processing and its Applications (DSPA), International Conference on

作者： Vladimir Bondarev Institute of Information Technologies Sevastopol State University Sevastopol Russian Federation

The article deals with signal processing models that use the representation of signals by spike trains. The “input-output” model of an integrate and fire spiking neuron is considered. The possibilities of using this model for spectral analysis and adaptive filtering are investigated. To solve the problem of adaptive processing of spike sequences, a generalized multi-input vector-matrix model of a spiking neuron is introduced. Based on this model, we derive supervised learning rules to adapt the dynamics of the spiking neuron. Examples of adaptive signal recovery from spike trains, additive noise suppression, and adaptive synthesis of band-pass filters are demonstrated.

关键词： Adaptation models Deconvolution Filtering Computational modeling Neurons Supervised learning signal processing algorithms

来源：评论

学校读者我要写书评

暂无评论

Large Direction-of-Arrival Mismatch Correction for Adaptive Beamforming

引用

IEEE ACCESS 2022年 10卷 61201-61212页

作者： Chang, Lin Yang, Hua Gulliver, T. Aaron Tan, Shizhe Wang, Yan Ocean Univ China Dept Elect Engn Qingdao 266100 Peoples R China Univ Victoria Dept Elect & Comp Engn Victoria BC V8W 2Y2 Canada Taishan Univ Dept Phys & Elect Engn Tai An 271000 Shandong Peoples R China

Direction-of-arrival (DOA) mismatch can degrade the performance of adaptive beamforming algorithms. Thus, a projection method is proposed to correct this mismatch. In a beamforming algorithm, the DOA error is usually regarded as a steering vector error which is corrected using a steering vector optimization algorithm. This approach can provide an optimal steering vector but ignores the actual DOA estimate. The proposed algorithm provides correction after DOA estimation but before beamforming to improve both the DOA estimation accuracy and beamforming gain. First, the signal-to-noise ratio (SNR) of the signal is estimated and used to regularize the covariance matrix. Then, an estimated steering vector with DOA close to the true value is determined based on a minimum number of projections. Numerical results are presented to verify the effectiveness of the proposed method for DOA estimation correction. In most cases, this method improves the performance of the beamforming algorithms without changing them.

关键词： Direction-of-arrival estimation Array signal processing signal to noise ratio Estimation Covariance matrices Interference signal processing algorithms Direction-of-arrival (DOA) mismatch correction adaptive beamforming covariance matrix regularization projection method

来源：评论

学校读者我要写书评

暂无评论

Deep Optimization of Parametric IIR Filters for Audio Equalization

引用

IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE processing 2022年 30卷 1136-1149页

作者： Pepe, Giovanni Gabrielli, Leonardo Squartini, Stefano Tripodi, Carlo Strozzi, Nicolo Univ Politecn Marche Dept Informat Engn Via Brecce Blanche I-60131 Ancona AN Italy ASK Ind SPA Viale Ramazni I-42124 Reggio Emilia Italy Univ Politecn Marche Dept Informat Engn Via Brecce Bianche I-60131 Ancona Italy

This paper describes a novel Deep Learning method for the design of IIR parametric filters for automatic multipoint audio equalization, that is the task of improving the sound quality of a listening environment at multiple listening points employing multiple loudspeakers. The filters are designed to approximate the inverse of the RIR and achieve almost flat magnitude response. A simple and effective neural architecture, named BiasNet, is proposed to determine the IIR equalizer parameters. This novel architecture is conceived for optimization and, as such, is able to produce optimal IIR equalizer parameters at its output, after training, with no input required. In absence of input, the presence of learnable non-zero bias terms ensures that the network works properly. An output scaling method is used to obtain accurate tuning of the IIR filters center frequency, quality factor and gain. All layers involved in the proposed method are shown to be differentiable, allowing backpropagation to optimize the network weights and achieve, after a number of training iterations, the optimal output according to a given RIR. The parameters are optimized with respect to a loss function based on a spectral distance between the measured and desired magnitude response, and a regularization term is used to keep the same microphone-loudspeaker energy balance after equalization. Two experimental scenarios are employed, a room and a car cabin, with several loudspeakers. The performance of the proposed method improves over the baseline techniques and achieves an almost flat band at a lower computational cost.

关键词： Equalizers Optimization Neural networks Speech processing signal processing algorithms MIMO communication Finite impulse response filters Deep learning optimization audio equalization machine learning IIR filtering parametric equalizer inverse filtering

来源：评论

学校读者我要写书评

暂无评论

2D-DOA Estimation in Arc-Array With a DNN Based Covariance Matrix Completion Strategy

引用

IEEE ACCESS 2022年 10卷 57608-57620页

作者： Tian, Ye Mei, Ruru Huang, Yonghui Tang, Xiaogang Cui, Tianshu Chinese Acad Sci Natl Space Sci Ctr Beijing 100190 Peoples R China Univ Chinese Acad Sci Beijing 100190 Peoples R China Space Engn Univ Sch Aerosp Informat Beijing 101416 Peoples R China Tsinghua Univ Beijing Natl Res Ctr Informat Sci & Technol Beijing 100084 Peoples R China

Two-dimensional direction of arrival (2D-DOA) estimation, or estimating the azimuth and elevation angles of sources simultaneously, is an increasingly important area in array signal processing. This paper presents a 2D-DOA algorithm together with a novel structure of array named Arc-Array (ArcA). The ArcA is enlarged to a virtual Uniform Circular Array (UCA) through a Deep-Neural-Network (DNN) based covariance matrix completion strategy;afterward, the MUSIC algorithm is performed with the completed covariance matrix. The proposed method is named as ArcA-DNN, and the performance of ArcA-DNN is evaluated by computer simulations. The simulation results indicate that the performance of 2D-DOA estimation in ArcA is able to approach that of a complete UCA;meanwhile, the number of physical elements is substantially reduced compared to the UCA. Moreover, the proposed ArcA-DNN algorithm gives access to implementing underdetermined 2D-DOA estimation with reasonable results.

关键词： Covariance matrices Estimation Antenna arrays Array signal processing signal processing algorithms Direction-of-arrival estimation Neurons 2D-DOA estimation UCA ArcA covariance matrix completion deep learning

来源：评论

学校读者我要写书评

暂无评论

Double Nonstationarity: Blind Extraction of Independent Nonstationary Vector/Component From Nonstationary Mixtures-algorithms

引用

IEEE TRANSACTIONS ON signal processing 2022年 70卷 5102-5116页

作者： Koldovsky, Zbynek Kautsky, Vaclav Tichavsky, Petr Tech Univ Liberec Fac Mechatron Proc GrpInformat Acoust Signal AnalInterdisciplinary Studies Liberec 46117 Czech Republic Czech Acad Sci Inst Informat Theory & Automat Prague 18208 Czech Republic

In this article, nonstationary mixing and source models are combined for developing new fast and accurate algorithms for Independent Component or Vector Extraction (ICE/IVE), one of which stands for a new extension of the well-known FastICA. This model allows for a moving source-of-interest (SOI) whose distribution on short intervals can be (non-)circular (non-)Gaussian. A particular Gaussian source model assuming tridiagonal covariance matrix structures is proposed. It is shown to be beneficial in the frequency-domain speaker extraction problem. The algorithms are verified in simulations. In comparison to the state-of-the-art algorithms, they show superior performance in terms of convergence speed and extraction accuracy.

关键词： signal processing algorithms Mathematical models Probability density function Covariance matrices Computational modeling Blind source separation Sensors Blind source extraction blind source separation dynamic models independent component analysis independent vector analysis moving sources

来源：评论

学校读者我要写书评

暂无评论

A Music Classification Approach Based on the Trajectory of Fifths

引用

IEEE ACCESS 2022年 10卷 73494-73502页

作者： Lukaszewicz, Tomasz Kania, Dariusz Silesian Tech Univ Fac Automat Control Elect & Comp Sci PL-44100 Gliwice Poland

In this paper we examine the applicability of the trajectory of fifths as a source of knowledge in automated music classification processes. The study shows that such trajectories provide valuable information concerning the harmonic structure of a given piece of music. The results of the conducted experiments indicate that even basic coefficients quantifying the trajectory of fifths (e.g. the length of the trajectory) can be utilized as feature variables in music classification algorithms. The performed experiments involved the computation of the trajectories of fifths for the music pieces from two genre groups - rock/pop and jazz, calculation of their basic coefficients, and using these coefficients as feature variables for various rock/pop - jazz classifiers based on popular machine learning algorithms, i.e. Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbors, Linear Discriminant Analysis, Gaussian Naive Bayes, and C-SVM. The tests evaluating the performance of the established classifiers have shown that in all of the considered cases the mean balanced accuracies were greater than 0.900, reaching the maximum of 0.936 (std=0.019) for the C-SVM classifier and the minimum of 0.910 (std=0.017) for the classification using the Decision Tree algorithm. Given the basic nature of the utilized coefficients and the relative simplicity of the applied classification models, the achieved results seem very promising. They strongly encourage further research concerning the use of the trajectory of fifths in different music classification algorithms - especially evaluation of its potential in the multi-class or multi-label genre classification scenarios.

关键词： Trajectory Music Classification algorithms Harmonic analysis Aggregates signal processing algorithms Instruments Music classification computer music analysis music information retrieval music data mining

来源：评论

学校读者我要写书评

暂无评论

3-D Array Image Data Completion by Tensor Decomposition and Nonconvex Regularization Approach

引用

IEEE TRANSACTIONS ON signal processing 2022年 70卷 4291-4304页

作者： Yang, Ming Luo, Qilun Li, Wen Xiao, Mingqing Southern Illinois Univ Carbondale Dept Comp Sci Carbondale IL 62901 USA Southern Illinois Univ Carbondale Dept Math Carbondale IL 62901 USA South China Normal Univ Sch Math Sci Guangzhou 510631 Guangdong Peoples R China

Various image datasets appear naturally in the form of multi-dimensional arrays (hypermatrices), called tensors. Image with incomplete entries, which often can be formulated as the low-rank tensor completion problem, is practically important in order to process large size 3D array images both efficiently and effectively. To make the low-rank approximation being tractable, most current approaches make use of the lower convex envelope of tensor multi-rank function for the approximation. Due to the gap between the rank function and its lower convex envelope, the adoption of tensor nuclear norm may lead to the approximation of the corresponding tensor tubal-rank being insufficient. In this paper, we introduce a new nonconvex regularization approach, which can better capture the low-rank characteristics than the convex approach. By transforming the original problem to the Fourier domain, we formulate an equivalent optimization problem with more transparent tensor rank characteristics, whose explicit solution can be obtained under our framework. A minimization algorithm, associated with the augmented Lagrangian multipliers and the nonconvex regularizer, is established and is shown to be feasible. The constructed sequence converges to the desirable Karush-Kuhn-Tucker point, which is mathematically validated in detail. Extensive experimental results demonstrate that our proposed approach outperforms the existing state-of-the-art convex approaches consistently.

关键词： Tensors Optimization Convergence Approximation algorithms signal processing algorithms Minimization Indexes Compressive sensing tensor completion low rank approximation non-convex regularization

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：