Error modeling of 3-RSR parallel robot based on D-H transformation matrix

By selecting any one limb of 3-RSR parallel robot as a research object, the paper establishes a position and orienta- tion relationship matrix between the moving platform and the base by means of Denavit-Hartenberg （D-H） transformation matrix. The error mapping model is derived from original error to the error of the platform by using matrix differential method. This model contains all geometric original errors of the robot. The nonlinear implicit function relation between po- sition and orientation error of the platform and the original geometric errors is simplified as a linear explicit function rela- tion. The results provide a basis for further studying error analysis and error compensation.

基于Transformer紧凑编码的局部近重复视频检测算法

针对现有局部近重复视频检测算法特征存储消耗大、整体查询效率低、提取特征时并未考虑近重复帧之间细微的语义差异等问题,文中提出了一种基于Transformer紧凑编码的局部近重复视频检测算法。首先,提出了一个基于Transformer的特征编码器,其学习了大量近重复帧之间细微的语义差异,可以在编码帧特征时对各个区域特征图引入自注意力机制,在有效降低帧特征维度的同时也提高了编码后特征的表示性。该特征编码器通过孪生网络训练得到,该网络不需要负样本就可以有效学习近重复帧之间的相似语义信息,因此无需沉重和困难的难负样本标注工作,使得训练过程更加简易和高效。其次,提出了一个基于视频自相似度矩阵的关键帧提取方法,可以从视频中提取丰富但不冗余的关键帧,从而使关键帧特征序列能够更全面地描述原视频内容,提升算法的性能,同时也大幅减少了存储和计算冗余关键帧带来的开销。最后,基于关键帧的低维紧凑编码特征,采用基于图网络的时间对齐算法,实现局部近重复视频片段的检测和定位。该算法在公开的局部近重复视频检测数据集VCDB上取得了优于现有算法的实验性能。

基于CNN-Transformer模型的自动扶梯轴承故障诊断方法研究

针对自动扶梯滚动轴承故障诊断模型诊断识别率不高的问题,提出一种CNN与Transformer相结合的滚动轴承故障诊断方法。该方法采用CNN注意力和Transformer时域注意力融合机制,从信号的通道和时域两个维度提取最有利于识别的特征信息,降低信道和噪声等影响。根据特征信息的重要程度进行权重分配,以提升故障识别率。基于滚动轴承数据集,将所提算法模型和CNN、RNN、SVM算法进行对比。结果表明,该方法识别轴承故障的准确度更高,实现了对轴承故障的准确分类。

Transformer与CNN并行引导的水下图像增强

为克服水下图像对比度低和色偏的问题,提出了基于Transformer与CNN并行引导的水下图像增强算法。利用3D位置嵌入模型为Transformer提供相对位置信息、色偏信息和特征图的全局特征,利用CNN编码器提取图像局部特征,将Transformer提取的全局特征和CNN提取的局部特征通过特征调制矩阵整合在一起,通过CNN解码器提高图像的分辨率,将解码器输出的特征图输入到特征加强网络中,由特征加强网络输出最终结果。采用现有的EUVP配对数据集进行训练,为验证该算法的优越性,选取具有不同程度色偏的水下图像进行定性比较和定量实验,结果显示,该算法增强后的水下图像峰值信噪比指标(peak signal-to-noise ratio,PSNR)和结构相似性指标(structural similarity index measure,SSIM)均高于其他对比算法,主观质量也得到显著提高,能够产生颜色丰富且清晰度较高的增强图像。

Applications of Lodrigues Matrix in 3D Coordinate Transformation

Three transformation models （Bursa-Wolf, Molodensky, and WTUSM） are generally used between two data systems transformation. The linear models are used when the rotation angles are small; however, when the rotation angles get bigger, model errors will be produced. In this paper, we present a method with three main terms：① the traditional rotation angles θ,φ,ψ are substituted with a,b,c which are three respective values in the anti-symmetrical or Lodrigues matrix; ② directly and accurately calculating the formula of seven parameters in any value of rotation angles; and ③ a corresponding adjustment model is established. This method does not use the triangle function. Instead it uses addition, subtraction, multiplication and division, and the complexity of the equation is reduced, making the calculation easy and quick.

CRACK IDENTIFICATION IN STEPPED CANTILEVER BEAM COMBINING WAVELET ANALYSIS WITH TRANSFORM MATRIX

This paper illustrates the crack identification method combining wavelet analysis with transform matrix. Firstly, the fundamental vibration mode was applied to wavelet analysis. The crack location was found by the peaks of the wavelet coefficients. Secondly, based on the identified crack locations, a simple transform matrix method requiring only the first two tested natural frequencies was used to further identify the crack depth. The present method can be used for crack identification in a complex structure. Numerical results of crack identification of a stepped cantilever beam show that the suggested method is feasible.

Effects of transforming growth factor β2 and connective tissue growth factor on induction of epithelial mesenchymal transition and extracellular matrix synthesis in human lens epithelial cells

AIM:To Investigate the effects of transforming growth factorβ2(TGF-β2)and connective tissue growth factor(CTGF)on transdifferentiation of human lens epithelial cells(HLECs)cultured in vitro and synthesis of extracellular matrix(ECM).METHODS:HLECs were treated with TGF-β2(0,0.5,1.0,5,10μg/L)and CTGF(0,15,30,60,100μg/L)for different times(0,24,48,72h)in vitro and the expression ofα-smooth muscle actin(α-SMA),the main component of the extracellular matrix typeⅠcollagen(Col-1)and fibronectin(Fn)were measured by using real-time polymerase chain reaction(PCR)and western-blot.RESULTS:TGF-β2 and CTGF significantly increased expression ofα-SMA mRNA and protein(P【0.05,P【0.001),Fn mRNA and protein(P【0.001),Col-1 mRNA and protein(P【0.001).TGF-β2 could induce HLECs expression of CTGF mRNA and protein in dosedependent manner(P【0.05,P【0.001).TGF-β2 and CTGF could induce HLECs to expressα-SMA,Fn and Col-1 in time-dependent manner.Each time of TGF-β2and CTGF induced HELCs expression ofα-SMA,Fn,Col-1 mRNA and protein was significant increase compared with control(P【0.05,P【0.001).CONCLUSION:TGF-β2 and CTGF could induce HLECs epithelial mesenchymal transition and ECM synthesis.

Image Fusion Based on Complex Contourlet Transform and Nonnegative Matrix Factorization

An image fusion method combining complex contourlet transform(CCT) with nonnegative matrix factorization(NMF) is proposed in this paper.After two images are decomposed by CCT,NMF is applied to their highand low-frequency components,respectively,and finally an image is synthesized.Subjective-visual-quality of the image fusion result is compared with those of the image fusion methods based on NMF and the combination of wavelet /contourlet /nonsubsampled contourlet with NMF.The experimental results are evaluated quantitatively,and the running time is also contrasted.It is shown that the proposed image fusion method can gain larger information entropy,standard deviation and mean gradient,which means that it can better integrate featured information from all source images,avoid background noise and promote space clearness in the fusion image effectively.

Determination of Transmission Routing of OXC Nodes Based on AWG Multiplexer by Matrix Transformation

The structures of the space switching and the wavelength switching optical cross connect (OXC) nodes which are based on the arrayed waveguide grating (AWG) multiplexer are analyzed.By the matrix transformation relation between the input and output wavelengths of the AWG multiplexer, the wavelength transmission routings of the space switching and wavelength switching OXC nodes are determined.

Inverse Transformation of Elliptical Relative State Transition Matrix

A new set of relative orbit elements (ROEs) is used to derive a new elliptical formation flying model in previous work. In-plane and out-of-plane relative motions can be completely decoupled, which benefits elliptical formation design. In order to study the elliptical control strategy and perturbation effects, it is necessary to derive the inverse transformation of the relative state transition matrix based on relative orbit elements. Poisson bracket theory is used to obtain the linear transformations between the two representations: the relative orbit elements and the geocentric orbital frame. In this paper, the details of these transformations are presented.

BANACH SPACES AND INEQUALITIES ASSOCIATED WITH NEW GENERALIZATION OF CESàRO MATRIX

Let the triangle matrix A^(ru)be a generalization of the Cesàro matrix and U∈{c_(0),c,l_(∞)}.In this study,we essentially deal with the space U(A^(ru))defined by the domain of A^(ru)in the space U and give the bases,and determine the Kothe-Toeplitz,generalized K?theToeplitz and bounded-duals of the space U(A^(ru)).We characterize the classes(l_(∞)(A^(ru)):l_(∞)),(l_(∞)(A^(ru)):c),(c(A^(ru)):c),and(U:V(A^(ru)))of infinite matrices,where V denotes any given sequence space.Additionally,we also present a Steinhaus type theorem.As an another result of this study,we investigate the l_(p)-norm of the matrix A^(ru)and as a result obtaining a generalized version of Hardy's inequality,and some inclusion relations.Moreover,we compute the norm of well-known operators on the matrix domain l_(p)(A^(ru)).

Effects of Heparin on Transforming Growth Factor-β_1 and Extracellular Matrix Components in the Glomeruli of Diabetic Rats

The effects of heparin on the expression of transforming growth factor-β 1 (TGF-β 1) and two extracellular matrix components laminin (LN) and fibronectin (FN) in diabetic rat glomeruli were investigated. Twenty-six rats were randomly divided into control group (C, n=8), diabetic group (D, n=9), and diabetes+heparin group (DH, n=9). After 8-week therapy of heparin (200 U once daily by abdominal injection), TGF-β 1, LN and FN expression in glomeruli was detected by immunohistochemical method. The results showed that the expression levels of TGF-β 1, LN and FN were higher in group D than in group C. It was found that heparin could reduce 24-h urinary albumin excretion and inhibit overexpression of TGF-β 1, LN and FN in glomeruli of diabetic rats. It suggested that the inhibitory effect of heparin on diabetic glomerular sclerosis was at least partly related with the inhibition of TGF-β 1 expression.

Matrix transformations of l^q(X) to l^p(Y )

Let X and Y be Banach spaces, 0 〈 q 〈 ＋∞, i ≤ p 〈 ＋∞. In this paper, we characterize matrix transformations of lq （ X ） to lp （ Y ）.

A Direct Transformation of a Matrix Spectrum

A method is presented for calculating a matrix spectrum with a given set of eigenvalues. It can be used to build systems with different spectrums with the aim of choosing desired alternative. It enables a practical implementation of control algorithms without resorting to transformation of variables.

Matrix Expression of the Orthogonal Wavelet (Packets) Transform

Matrix expression of finite orthogonal wavelet transform of finite impulse response signal is more valuable for theoretical analysis and understanding. However, clear deduction for matrix expression has not been provided yet. In this paper, the formulation to generate the re-lated matrix is put forward and the theorem on the orthogonality of this matrix proved. This effort deploys a basis for more deeper and wider applications in chemical processes. *

基于Dlib和变种Transformer的哈欠检测方法

为及时检测驾驶员的疲劳驾驶情况,提出了一种基于Dlib和变种Transformer的哈欠检测方法。首先,基于Dlib的人脸关键点模型构建驾驶员眼部和嘴部的哈欠特征矩阵,然后在视频检测领域提出一种变种Transformer模型对哈欠特征矩阵进行提取和分类,最后基于YawDD数据集进行验证,结果表明,所提出算法的哈欠检测准确率达96.8%,高于已有算法,适用于驾驶员疲劳驾驶时哈欠行为的检测。

结合Transformer的轻量化中文语音识别

近年来,深度神经网络模型在语音识别领域成为热门研究对象。然而,深层神经网络的构建依赖庞大的参数和计算开销,过大的模型体积也增加了其在边缘设备上部署的难度。针对上述问题,提出了基于Transformer的轻量化语音识别模型。首先使用深度可分离卷积获得音频特征信息;其次构建了双半步剩余权重前馈神经网络,即Macaron-Net结构,并引入低秩矩阵分解,实现了模型压缩;最后使用稀疏注意力机制,提升了模型的训练速度和解码速度。为了验证模型,在Aishell-1和aidatatang_200zh数据集上进行了测试。实验结果显示,与Open-Transformer相比,所提模型在字错误率上相对下降了19.8%,在实时率上相对下降了32.1%。

Research on the Applications of Probability Model and Matrix Transformation in Data Analysis and Management under the Environment of Information Time

In this paper, we research the probability theory and matrix transformation based technique to manage the data for processing and analysis. Clustering analysis research has a long history, over the decades, the importance and the cross characteristics with other research direction to get the affirmation of the people. The probability theory and linear algebra act as the powerful tool for analyzing and mining data. The experimental result illustrates the effectiveness. In the near future, we plan to conduct more theoretical analysis on the topic.

Transformation Matrix for Combined Loads Applied to Thin-Walled Structures

This paper transforms combined loads, applied at an arbitrary point of a thin-walled open section beam, to the shear centre of the cross-section of the beam. Therein, a generalized transformation matrix for loads with respect to the shear centre is derived, this accounting for the bimoments that develop due to the way the combined loads are applied. This and the authors’ earlier paper (World Journal of Mechanics 2021, 11, 205-236) provide a full solution to the theory of thin-walled, open-section structures bearing combined loading. The earlier work identified arbitrary loading with the section’s area properties that are necessary to axial and shear stress calculations within the structure’s thin walls. In the previous paper attention is paid to the relevant axes of loading and to the transformations of loading required between axes for stress calculations arising from tension/compression, bending, torsion and shear. The derivation of the general transformation matrix applies to all types of loadings including, axial tensile and compression forces, transverse shear, longitudinal bending. One application, representing all these load cases, is given of a simple channel cantilever with an eccentrically located end load.

Research on the File Encryption and Programming based on Matrix Transformation

Any linear transform matrix can be used to easily calculate a consistent form, and a plurality of conversion can be easily connected together by matrix multiplication. When performing file transfers, you can encrypt files matrix transformation. Article presents a matrix-based electronic document encryption and decryption algorithm, which relies on a special class of matrices combinatorial problems, the method to improve the security of electronic document system is feasible and effective, and finally give the source code and programming software.