基于压缩传感理论的绝缘子泄漏电流数据压缩

绝缘子泄漏电流数据量大,给监测系统数据存储和传输带来巨大的负担。针对这个问题,提出基于压缩传感理论(Compressed Sensing,CS)的数据压缩方法,CS将采样与压缩合并进行,少量采样就能很好地恢复信号,不仅降低对硬件要求,而且提高压缩效率。将泄漏电流信号进行小波变换的稀疏分解,然后对稀疏的泄漏电流信号进行高斯测量编码,最后应用正交匹配追踪算法(OMP)重构信号。实验结果表明,对绝缘子泄漏电流进行CS数据压缩具很高的压缩比,恢复的信号也比较理想。

压缩传感理论、优化算法及其在系统状态重构中应用

分别对压缩传感理论、优化算法及其在系统状态重构中的应用3个方面进行了研究.在压缩传感理论方面,包括对所压缩信号的稀疏或低秩要求、编码测量以及与优化算法之间的关系进行了较为深入的研究,重点分析了原始信号的稀疏与低秩之间的关系、测量矩阵与压缩矩阵之间的关系、满足限制等距特性(RIP)的测量矩阵,以及由压缩传感理论提供的最少测量次数.在压缩信号重构过程中所需要采用的优化算法,着重讨论了核函数的凸优化问题描述,分别对常用的优化算法,包括最小二乘(LS)法、最大熵法、极大似然法和贝叶斯方法的求解过程中所用到的性能指标、优化目标和求解条件等进行了归纳与特性分析.对量子态估计中的交替方向乘子法(ADMM)以及作者最新提出的迭代阈值收缩法(IST)进行了专门的性能对比,并通过量子位分别5、6和7情况下纯态估计的应用为例,对不同测量比率对参数估计性能的影响,以及算法在不同量子位数下性能的表现,进行了不同层次上的对比和分析,完整地阐述基于压缩传感理论与优化的系统参数估计的研究过程.

压电石英晶体生物传感器及其应用研究进展

压电石英晶体生物传感器是利用压电石英晶体振荡频率对晶体表面质量负载和表面性状如密度、粘度、电导、介电常数等的高度敏感性与生物识别分子的高度特异性相结合发展起来的一种新型传感器。它具有灵敏度高、特异性好 ;操作简单 ,不需任何标记 ;检测速度快 ,成本低廉 ;仪器体积小 ,重量轻 ,能实时检测等特点。特别适用于临床实验诊断、环境监测、食品卫生检验、工业生产实时监测等野外流动作业和在线检测。本文就压电石英晶体生物传感器的原理、基本结构、特点及其应用等方面进行综述。

基于压缩感知和离散余弦变换的数字图像水印算法

为了使水印信息更好地隐藏于原始图像中,加快水印信息检测,设计一种基于压缩感知和离散余弦变换的数字图像水印算法.首先将原始载体图像划分成多个子块,采用离散余弦变换对各子块进行分解;然后采用压缩感知对水印信息进行处理,减少嵌入的信息量;最后将水印信息嵌入到原始载体子块的相应位置,并通过匹配跟踪算法实现水印信息的检测.从主观和客观两方面对数字图像水印算法的性能进行评价,结果表明,该算法的水印具有良好的不可感知性,对抵抗常见攻击具有很好的鲁棒性,解决了水印透明性与鲁棒性之间的矛盾,且加快了水印提取速度.

应用CS理论实现同步采样压缩成像

为了减轻图像数据存储负担,实现图像在网络上的快速传输和实时处理,对一种新的压缩传感(CS)理论进行了研究。介绍了压缩传感理论的主要思想和基于压缩传感理论的光学成像系统,给出了一种新型图像重建算法—和谐正交匹配追踪算法,并进行了相应的模拟实验。实验结果显示,该成像机制可同步完成图像的采样与数据压缩,同时可获得良好的图像重建效果。由于该方法所要传输的信号数据量较小,所以十分有利于远距离的图像传输。

一种单检测器可压缩成像系统设计

采用基于亚高斯随机投影的图像重建方法,得到以稀疏矩阵、非常稀疏投影矩阵作为测量矩阵的仿真结果,设计一种基于数字微镜装置阵列的可压缩成像系统,给出系统结构、各模块之间的联系和核心模块的设计方法。为满足系统对高频弱光信号检测的需要,设计单检测器弱光信号检测模块,实验结果证明该方案速度快、有较高的精度和适应性。

改进的构建Wi-Fi位置指纹库算法研究

针对传统的位置指纹算法在更新位置指纹库时人力和物力巨大耗费的问题,提出利用压缩传感理论和重心拉格朗日插值算法来更新位置指纹库.压缩传感理论将指纹向量的重构过程转换为一个最小l0范数的优化问题,并通过最小全变分方法求解原始指纹向量.重心拉格朗日插值算法利用样本节点间的空间相关性,使得在离线阶段通过测量少量指纹就可重建位置指纹库.在真实室内环境的实验验证了压缩传感恢复算法比重心拉格朗日插值算法具有更好的定位性能.

基于EMD分解与压缩传感的泄漏电流压缩方法

为了提高绝缘子监测系统数据传输和处理速度,对非平稳泄漏电流信号进行数据压缩至关重要。压缩传感理论(Compressed Sensing,CS)将采样与压缩合并进行,少量采样就能很好地恢复信号,不仅降低对硬件要求,而且提高压缩效率。然而CS恢复非平稳信号过程中计算复杂度很高。提出经验模态分解(EMD)和CS相结合的方法,将非平稳泄漏电流分解成有限个相对平稳的固有模态IMF,然后对相对平稳的IMF分量进行CS处理。实验结果表明,CS处理经EMD分解的泄漏电流信号不仅提高整个系统的处理速度和运行效率,而且提高了压缩比和重构精度。

单量子比特系统状态的在线估计

针对具有退相干效应与测量反馈随机噪声的随机开放量子系统,采用对状态影响较弱的连续弱测量在线获取一系列状态的部分信息,实现量子状态的在线估计.由泡利矩阵构造初始测量算符,并推导出在线的随时间变化的测量算符;基于压缩传感理论来减少测量次数;采用最小二乘优化算法对自由演化中的量子密度矩阵状态进行重构,完整地给出了量子态在线估计的过程.所提出的在线量子态估计方案,在一个量子位系统上进行了系统仿真实验.数值仿真实验结果表明,在满足压缩传感理论的条件下,仅需2次连续弱测量所得到的测量值之后,就可以高精度地实现在线变化的单比特量子密度矩阵估计.

Cutting force measurement based on tool embedded Ni-chrome thin-film micro-sensors

Cutting force measurement has become a crucial activity for enhancing machining process performance. This paper described the design and fabrication of embedded Ni-chrome thin-film micro-sensors in tool holders to measure the cutting force in machining operations. A Ni-chrome thin-film sensor device is embedded within a substrate structure through a dy- namic brazing process, which consists of a Ti6A14V substrate, a nickel-chromium thin-film sensor and an alumina insulating layer. The Wheatstone bridge which consists of four sensors would produce the output voltage when the thin film caused de- formation by the cutting forces. The relationship between input and output voltages was theoretically analyzed. According- ly, an in-process cutting force measurement system is established. The results show that the thin-film sensor had good lineari- ty and less mutual interference, and it is suitable for all kinds of turning forces under the measurement conditions.

DEVELOPMENT OF EXTRINSIC FABRY PEROT FIBER SENSOR AND ITS APPLICATION TO SMART MATERIALS

In the work of developing extrinsic fabry perot interferometric (EFPI), the key technology of polishing fiber optic endfaces and coating the multilayer of dielectric films on them is raised and resolved to settle the disturbance and stability problem of EFPI, which simplifies the sensing system, improves the sensor performance and reduces the cost. In this paper, the relations between the output interferential light intensity and the F P cavity length are calculated based on the theory of mode field coupling. The EFPI fiber optic sensor is adhered to a distributed smart laminate beam to detect vibration frequency and axial strain value, the results coincident with the results tested by PZT.

TDDFT Study on Different Sensing Mechanisms of Similar Cyanide Sensors Based on Michael Addition Reaction

The solvents and substituents of two similar fluorescent sensors for cyanide, 7-diethylamino- 3-formylcoumarin （sensor a） and 7-diethylamino-3-（2-nitrovinyl）coumarin （sensor b）, are proposed to account for their distinct sensing mechanisms and experimental phenomena. The time-dependent density functional theory has been applied to investigate the ground states and the first singlet excited electronic states of the sensor as well as their possible Michael reaction products with cyanide, with a view to monitoring their geometries and photophysieal properties. The theoretical study indicates that the protic water solvent could lead to final Michael addition product of sensor a in the ground state, while the aprotic acetonitrile solvent could lead to carbanion as the final product of sensor b. Furthermore, the Michael reaction product of sensor a has been proved to have a torsion structure in its first singlet excited state. Correspondingly, sensor b also has a torsion structure around the nitrovinyl moiety in its first singlet excited state, while not in its carbanion structure. This could explain the observed strong fluorescence for sensor a and the quenching fluorescence for the sensor b upon the addition of the cyanide anions in the relevant sensing mechanisms.

Compressive Sensing Based Wireless Localization in Indoor Scenarios

The sparse nature of location finding in the spatial domain makes it possible to exploit the Compressive Sensing (CS) theory for wireless location.CS-based location algorithm can largely reduce the number of online measurements while achieving a high level of localization accuracy,which makes the CS-based solution very attractive for indoor positioning.However,CS theory offers exact deterministic recovery of the sparse or compressible signals under two basic restriction conditions of sparsity and incoherence.In order to achieve a good recovery performance of sparse signals,CS-based solution needs to construct an efficient CS model.The model must satisfy the practical application requirements as well as following theoretical restrictions.In this paper,we propose two novel CS-based location solutions based on two different points of view:the CS-based algorithm with raising-dimension pre-processing and the CS-based algorithm with Minor Component Analysis (MCA).Analytical studies and simulations indicate that the proposed novel schemes achieve much higher localization accuracy.

Suboptimal reliable guaranteed cost control for continuous-time systems with multi-criterion constraints

The suboptimal reliable guaranteed cost control （RGCC） with multi-criterion constraints is investigated for a class of uncertain continuous-time systems with sensor faults. A fauk model in sensors, which considers outage or partial degradation of sensors, is adopted. The influence of the disturbance on the quadratic stability of the closed-loop systems is analyzed. The reliable state-feedback controller is developed by a linear matrix inequalities （LMIs） approach, to minimize the upper bound of a quadratic cost fimction under the conditions that all the closed-loop poles be placed in a specified disk, and that the prescribed level of H∞ disturbance attenuation and the upper bound constraints of control inputs＇ magnitudes be guaranteed. Thus, with the above muki-criterion constraints, the resulting closed-loop system can provide satisfactory stability, transient property, a disturbance rejection level and minimized quadratic cost performance despite possible sensor faults.

Application of fiber Bragg grating sensor network in aluminum reduction tank shell temperature monitoring

A fiber Bragg grating temperature sensor network was designed to implement the real-time health monitoring of the aluminum reduction cell. The heat transfer process was simulated using software ANSYS, and an on-line shell monitoring system was established based on optical sensing technology. According to aluminum reduction cell heat transfer theory, the 2D slice finite element model was developed. The relationship between shell temperature and cell status was discussed. Fiber Bragg grating （FBG） was chosen as the temperature sensor in light of its unique advantages. The accuracy of designed FBG temperature sensors exceeds 2 ~C, and good repeatability was exhibited. An interrogation system with 104 sensors based on VPG （volume phase grating） filter was established. Through the long-term monitoring on running state, the status of the aluminum reduction cell, including security and fatigue life could be acquired and estimated exactly. The obtained results provide the foundation for the production status monitoring and fault diagnosis. Long-term test results show good stability and repeatability which are compatible with electrolysis process.

Laser triangulation measurement of the level in a coal silo

Laser triangulation theory was used to develop a novel contact-free method for measuring the coal level in a silo under harsh environmental conditions found in coal mines, such as the presence of dense dust, high humidity, and low illumination. A laser source and a camera were mounted at the top of the silo. The laser spot projected into the silo was imaged by the camera. The pinhole imaging principle allows the level to be found from the lateral shift of the spot image on the sensor. A pre-calibrated look-up table of the coal depth versus spot position was used to obtain the depth. The measurement accuracy depends on the step size used during pre-calibration. The actual application of a device designed according to these principles shows that it is easy to implement. The detection of the coal level in a silo at the low illumination level found in coal mines is demonstrated.

Design of ADS-B Simulator

In this paper, the ADS-B （automatic dependent surveillance-broadcast） used for manned aircrafts is chosen for aSee-and-Avoid sensor of UAV （unmanned aerial vehicle）, then GDL90 type is selected among three types of ADS-B. The serialcommunication data protocol which communicates between GDL90 and control panel of GDL90 is analyzed and tailored for theinterface of the GDL90 simulator here. This ADS-B （GDL90） simulator sends intruder information to FCC （flight control computer）of UAV for collision avoidance function.

Capacity Research in Cluster-Based Underwater Wireless Sensor Networks Based on Stochastic Geometry

Underwater wireless sensor networks(UWSNs) have attracted wide attention in recent years.The capacity research on it is still in the initial stage,lacking adequate performance evaluation for network construction.This paper will focus on this subject by theoretical analysis and simulation,aiming to provide some insights for the actual UWSNs construction.According to the structure features of cluster-based UWSNs and the propagation characteristics of underwater acoustic signal,with the combination of signal to interference plus noise ratio,we define some capacity performance metrics,such as outage probability and transmission capacity.Based on the theory of stochastic geometry,a network capacity analytical model used in the cluster-based UWSNs is presented.The simulation results verify the validity of the theoretical analysis,and the cause of error between theoretical and simulation results has also been clearly explained.

Electronic Response of Nano-sized Cages of ZnO and MgO to Presence of Nitric Oxide

We have performed a comparative theoretical study on the adsorption of nitric oxide （NO） on Zn12012 and Mgt2012 nanocages in terms of their energetic, geometric, and electronic properties. It has been found that NO adsorption on the MgO nanocage is energetically more favorable than that on the ZnO one. In contrast to the ZnO nanocage, HOMO-LUMO energy gap （Eg） of MgO one is dramatically decreased in the presence of NO molecule so that it is transformed from an intrinsic semiconductor （Eg≈5.00 eV） to a p-type one （Eg≈1.93 eV）. We have predicted that electronic and conductance properties of the Mg12012 nanocage are sensitive toward NO molecule, thus it may be potential candidate in detection of NO molecules.

Highly Effective Detection of Amitraz in Honey by Using Surface-Enhanced Raman Scattering Spectroscopy Coupled with Chemometric Methods

As an effective and universal acaricide, amitraz is widely used on beehives against varroasis caused by the mite Varroa jacobsoni. Its residues in honey pose a great danger to human health. In this study, a sensitive, rapid, and environmentally friendly surface-enhanced Raman spectroscopy method (SERS) was developed for the determination of trace amount of amitraz in honey with the use of silver nanorod (AgNR) array substrate. The AgNR array substrate fabricated by an oblique angle deposition technique exhibited an excellent SERS activity with an enhancement factor of -10^7. Density function theory was employed to assign the characteristic peak of amitraz. The detection of amitraz was further explored and amitraz in honey at concentrations as low as 0.08 mg/kg can be identified. Specifically, partial least square regression analysis was employed to correlate the SERS spectra in full-wavelength with Camitraz to afford a multiple-quantitative amitraz predicting model. Preliminary results show that the predicted concentrations of amitraz in honey samples are in good agreement with their real concentrations. Compared with the conventional univariate quantitative model based on single peak’s intensity, the proposed multiple-quantitative predicting model integrates all the characteristic peaks of amitraz, thus offering an improved detecting accuracy and anti-interference ability.