子带t分布的快速独立向量分析在语声盲源 分离中的应用*

为了提高独立向量分析算法在盲语声分离任务中的分离性能,降低算法计算复杂度,并改善目前尚未完全解决的顺序模糊性的问题,该文提出一种基于子带t分布的快速独立向量分析算法。在声源模型方面,该算法首先利用语声信号重尾分布的特性,假设声源概率密度函数服从t分布,同时采用子带建模的方法来增强同一声源相邻频点的高阶依赖性,进而减轻频点间的顺序不一致问题。在空间模型方面,该算法采用秩1更新的方式估计声源信号,避免矩阵求逆操作和分离矩阵的估计,从而降低计算复杂度。实验结果表明,与现有的基于独立向量分析的盲源分离算法相比,该算法能够在相同的迭代次数下取得更优的语声分离性能。     

任意层互连生产技术研究

任意层互连（ELIC）是最高阶的高密度连接（HDI）制程，它可以比传统线路在层数相同的情况下，增加约30%层间互连；随着移动通讯产品功能需求的增加，ELIC技术被大举应用在智能手机及平面电脑等产品上。高阶ELIC线路板的生产不但流程长，且生产难度较大。本文对ELIC板的制作难点进行了分析，并提出相应的解决方法和生产注意事项，为ELIC工艺产业化提供必要的参考依据。     

A robust adaptive weighted constant modulus algorithm for blind equalization of wireless communications systems under impulsive noise environment

A robust adaptive weighted constant modulus algorithm is proposed for blind equalization of wireless communication systems under impulsive noise environment. The influence of the impulsive noise is analyzed based on numerical analysis method. Then an adaptive weighted constant modulus algorithm is constructed to adaptively suppress impulsive noise. Theoretical analysis is provided to illustrate that the proposed algorithm has a robust equalization performance since the impulsive noise is adaptively suppressed. Moreover, the proposed algorithm has stable and quick convergence due to avoidance of large misadjuntment and adoption of large step size. Simulation results are presented to show the robust equalization performance and the fast convergence speed of the proposed algorithm under both impulsive noise and Gaussian noise environments.     

RGB-D image-based detection of stairs,pedestrian crosswalks and traffic signs

A computer vision-based wayfinding and navigation aid can improve the mobility of blind and visually impaired people to travel independently. In this paper, we develop a new framework to detect and recognize stairs, pedestrian crosswalks, and traffic signals based on RGB-D (Red, Green, Blue, and Depth) images. Since both stairs and pedestrian crosswalks are featured by a group of parallel lines, we first apply Hough transform to extract the concurrent parallel lines based on the RGB (Red, Green, and Blue) channels. Then, the Depth channel is employed to recognize pedestrian crosswalks and stairs. The detected stairs are further identified as stairs going up (upstairs) and stairs going down (downstairs). The distance between the camera and stairs is also estimated for blind users. Furthermore, the traffic signs of pedestrian crosswalks are recognized. The detection and recognition results on our collected datasets demonstrate the effectiveness and efficiency of our proposed framework.     

Site mapping and small molecule blind docking reveal a possible target site on the SARS-CoV-2 main protease dimer interface

The SARS-CoV-2 virus is causing COVID-19 resulting in an ongoing pandemic with serious health, social, and economic implications. Much research is focused in repurposing or identifying new small molecules which may interact with viral or host-cell molecular targets. An important SARS-CoV-2 target is the main protease (M^pro), and the peptidomimetic α-ketoamides represent prototypical experimental inhibitors. The protease is characterised by the dimerization of two monomers each which contains the catalytic dyad defined by Cys¹⁴⁵ and His⁴¹ residues (active site). Dimerization yields the functional homodimer. Here, our aim was to investigate small molecules, including lopinavir and ritonavir, α-ketoamide 13b, and ebselen, for their ability to interact with the M^pro. The sirtuin 1 agonist SRT1720 was also used in our analyses. Blind docking to each monomer individually indicated preferential binding of the ligands in the active site. Site-mapping of the dimeric protease indicated a highly reactive pocket in the dimerization region at the domain III apex. Blind docking consistently indicated a strong preference of ligand binding in domain III, away from the active site. Molecular dynamics simulations indicated that ligands docked both to the active site and in the dimerization region at the apex, formed relatively stable interactions. Overall, our findings do not obviate the superior potency with respect to inhibition of protease activity of covalently-linked inhibitors such as α-ketoamide 13b in the M^pro active site. Nevertheless, along with those from others, our findings highlight the importance of further characterisation of the M^pro active site and any potential allosteric sites.     

Blind velocities mitigation for MIMO GMTI radar with Doppler division multiple access waveforms

Multiple-input multiple-output (MIMO) radar with multiple transmitters and multiple receivers can achieve a larger virtual antenna array and more system degrees of freedom; thus applying it to ground moving target indication (GMTI) radar can improve the performance of GMTI. Doppler division multiple access (DDMA) waveforms are approximately orthogonal providing good minimum detectable velocity (MDV) performance. However, in such DDMA systems, a sufficient pulse repetition frequency (PRF) design freedom is required. Furthermore, these waveforms suffer from blind velocities which are serious problems, especially in radar systems with high carrier frequency or low PRF. This paper analyses the blind velocities problem and show that blind velocities are relative to variation of the PRF and/or the carrier frequency. Variable PRF techniques are widely used in conventional GMTI radar including multiple PRFs and variable pulse repetition intervals (PRI). Combined with the characteristics of the DDMA MIMO GMTI radar, this paper proposed two methods to mitigate blind velocities: “multi-PRF DDMA” which employs multiple PRFs over successive coherent processing intervals, and “PRI-dithered DDMA” which employs nonuniform sampling by dithered PRI in slow time. Simulation results demonstrate that both the methods are effective ways to mitigate blind velocities in DDMA MIMO GMTI radar systems.     

GSM蜂窝无线通信系统

ＧＳＭ数字移动通信系统是现今较先进的一种无线通信系统，本文对该系统的组成特点、功能原理、信令系统和帧结构，以及今后的发展等作了较详细的介绍，以飨读者。     

基于NPCA的盲源分离算法

主要讨论了基于非线性主分量分析(NPCA)的盲源分离,从理论与实验2个方面详细分析了算法的特性与效果。针对算法中的非线性函数选择的问题,采用了在线统计的方法,即根据不同的输入信号选择不同的非线性函数。从实验结果可以看出,该方法不仅可以很好地解决源信号为亚高斯信号混合的盲源分离问题,而且对源信号为亚高斯和超高斯信号混合的盲源分离问题也取得了很好的效果。     

基于盲源分离的多幅顺序图像加密方法

根据许多加密方法利用数学难题保障其安全性的思想,盲源分离欠定难题可用于高度安全的多幅图像加密。然而,由于盲源分离存在顺序和幅度模糊性,通过盲源分离得到的解密图像可能发生顺序变化和像素值反转。这在加密多幅顺序图像时可能导致解密错误。针对这一问题,该文利用数字水印技术,加密前在每幅明文图像中嵌入与其对应的顺序信息;解密后,通过在各解密图像中检测该顺序信息而消除其顺序和幅度的模糊性。计算机仿真结果表明,该方法在恢复解密图像顺序的同时也能检测其是否反转,从而有效地解决了多幅顺序图像的盲源分离加密问题。