Order Vector Research Articles

Robust Constrained MFMVDR Filters for Single-Channel Speech Enhancement Based on Spherical Uncertainty Set

Aiming at exploiting speech correlation across consecutive time-frames in the short-time Fourier transform domain, the multi-frame minimum variance distortionless response (MFMVDR) filter for single-channel speech enhancement has been proposed. The MFMVDR filter requires an accurate estimate of the normalized speech correlation vector in order to avoid speech distortion and artifacts. In this paper we investigate the potential of using robust MVDR filtering techniques to estimate the normalized speech correlation vector as the vector maximizing the total signal output power within a spherical uncertainty set, which corresponds to imposing a quadratic inequality constraint. Whereas the singly-constrained (SC) MFMVDR filter only considers the quadratic inequality constraint to estimate the (non-normalized) speech correlation vector, the doubly-constrained (DC) MFMVDR filter integrates a linear normalization constraint into the optimization problem to directly estimate the normalized speech correlation vector. To set the upper bound of the quadratic inequality constraint for each time-frequency point, we propose to use a trained non-linear mapping function that depends on the a-priori signal-to-noise ratio (SNR). Experimental results for different speech signals, noise types and SNRs show that the proposed constrained approaches yield a more accurate estimate of the normalized speech correlation vector than a state-of-the-art maximum-likelihood (ML) estimator. An instrumental and a perceptual evaluation show that both constrained MFMVDR filters lead to less speech and noise distortion but a lower noise reduction than the ML-MFMVDR filter, where the DC-MFMVDR filter is preferred in terms of overall quality compared to the SC-MFMVDR and ML-MFMVDR filters.

Virtual space vector modulation strategy for common‐mode voltage reduction in three‐phase three‐level flying‐capacitor split‐source inverters

Transformerless systems play an important role in many applications, such as photovoltaic generation systems and motor drives, since they present reduced size and high efficiency. On the other hand, the flying-capacitor split-source inverter (FC-SSI) merges the boost and the inverter stage in a single power conversion system. It is important to point out that conventional modulation strategies applied to the FC-SSI can result in unacceptable common-mode voltage (CMV) which is one of the main causes of leakage currents. To overcome this problem, this paper proposes a virtual space vector modulation technique for FC-SSI. The proposed modulation strategy manipulates the switching vectors in order to result in constant CMV enabling the transformerless operation. Simulation and experimental results, as well as a comparative analysis with existing modulation techniques, are provided to show the good features of the proposed modulation strategy.

A correction method of word spelling mistake for English text

Spell correction is already a mature field, the need to combine advantages of different methods for better performance arises in non-word problem. A combined spell correction method is proposed in this paper, which contains Levenshtein distance for comparation between misspelled words and correct spelled words in dictionary, improved Double Metaphone algorithm that includes vowel phoneme rule sets aimed at Chinese English learners, and global vectors(GloVe) for character representation that can generate vectors in order to obtain better suggestion lists for misspelled words. Result shows that the combined approach proposed in this paper is better than phonetic correction or edit distance method only, and a comparison with two generally implemented spell check tools is done for experiment which shows that this approach is better than them in correcting misspelled words, and the success rates of suggestion lists for spelling mistakes hit the spot.

Quenching of an antiferromagnet into high resistivity states using electrical or ultrashort optical pulses

Ultra-fast dynamics, insensitivity to external magnetic fields, or absence of magnetic stray fields are examples of properties that make antiferromagnets of potential use in the development of spintronic devices. Similar to their ferromagnetic counterparts, antiferromagnets can store information in the orientations of the collective magnetic order vector. However, also in analogy to ferromagnets, the readout magnetoresistivity signals in simple antiferromagnetic films have been weak and the extension of the electrical reorientation mechanism to optics has not been achieved. Here we report reversible and reproducible quenching of an antiferromagnetic CuMnAs film by either electrical or ultrashort optical pulses into nano-fragmented domain states. The resulting resistivity changes approach 20\% at room temperature, which is comparable to the giant magnetoresistance ratios in ferromagnetic multilayers. We also obtain a signal readout by optical reflectivity. The analog time-dependent switching and relaxation characteristics of our devices can mimic functionality of spiking neural network components.

Characterization of the sand fly fauna in Barbalha, one of the municipalities with the highest leishmaniasis rates in Brazil.

Leishmaniases are a complex of sand fly-borne diseases that are considered a public health issue in several countries. Brazil presents high leishmaniases rates. The South of Ceará State, known as Cariri region, shows worrying statistics mainly on American tegumentary leishmaniasis. In Barbalha, which is one of the municipalities in this region, there is still a lack of studies regarding the local phlebotomine (Diptera: Psychodidae) fauna in order to help clarify the high rates. This study aimed to characterize such fauna by capturing sand flies with light traps during a four-year period. A total of 3730 sand flies were captured, of which 37.8% were females. Fourteen species were found: 13 of the Lutzomyia genus and one of the Brumptomyia genus. Of the Lutzomyia species, four were proven and five had potential involvement in leishmaniasis transmission. Lutzomyia longipalpis was the most common species (66.97%). This predominance, especially in the urban area, indicates its epidemiological importance and adaptation to environmental conditions modified by human activity. In fact, further studies are still required to accurately determine the behavioral features of these vectors in order to guide public health measurements towards its control and prevention.

Polarization modulation of vector bisolitons

In this paper, we simulate the polarization modulation of vector bisolitons at 1-μm wavelength region in a polarization resolved optical fiber system. For the case of polarization-locked vector bisoliton, the change of projection angle and phase difference can result the change of time delay and peak intensity ratio of output vector bisolitons, respectively. While for the case of group-velocity-locked vector bisoliton, vector bisolitons with multiple peaks/dips can be generated through changing projection angle, time delay and phase difference. Besides, the third pulse will appear when time delay reaches a specific value for both cases, which means pseudo-high- order vector solitons are generated. This work can provide beneficial conduct for relevant experiments about out-cavity polarization modulation of vector bisolitons.

Optimization of the Receiving Orientation Angle for Zero-Forcing Precoding in VLC

We study the performance of linear zero-forcing (ZF) precoding in multiuser multiple-input single-output visible light communications (VLC) when we are able to select the receiving orientation angle (ROA) of each user. For radio-frequency communications, the non-line-of-sight rich scattering environment usually ensures the linear independence among user’s channels. However, this condition is less likely to happen in VLC systems, degrading the performance of ZF precoding. In this work, we propose a variable ROA (vROA) photodetector able to modify its orientation vector in order to generate semi-orthogonal channel responses among users. We derive the algorithm for determining the orientation of the vROA photodetector of each user, obtaining optimal and suboptimal solutions with high and low complexity, respectively. Simulation results show that the performance of ZF precoding is improved considerably by managing the users ROA.

Helical magnetic ordering studied in single-crystalline GdBe13

The beryllide ${\mathrm{GdBe}}_{13}$ with the ${\mathrm{NaZn}}_{13}$-type face-centered-cubic structure has been known to undergo a proper helical-magnet ordering from experimental studies using polycrystalline samples. In the present study, we carried out electrical resistivity, specific heat, and magnetization measurements of single-crystalline ${\mathrm{GdBe}}_{13}$ in order to investigate a mechanism of its helical ordering. These measurements reveal that the present compound is a metallic system exhibiting the magnetic ordering of local ${\mathrm{Gd}}^{3+}$ moments at ${T}_{\mathrm{N}}$ = 24.8 K accompanied with strong magnetic fluctuations extending to temperatures well above ${T}_{\mathrm{N}}$. Furthermore, we constructed a magnetic field--temperature ($B$--$T$) phase diagram for $B\phantom{\rule{4pt}{0ex}}\ensuremath{\parallel}$ [001]. It consists of a multidomain state, which is composed of magnetic structures with $B$ applied parallel and perpendicular to the helical plane, in the lower-magnetic-field region below $\ensuremath{\sim}0.45\phantom{\rule{0.16em}{0ex}}\mathrm{T}$ and a possible single-domain conical one in the higher-field region in the ordering state. The helical structure of ${\mathrm{GdBe}}_{13}$ characterized by an incommensurate ordering vector ${\mathbit{q}}_{0}$ of (0, 0, 0.285) is discussed on the basis of a competition of Heisenberg exchange interactions between the ${\mathrm{Gd}}^{3+}$ moments assuming an one-dimensional layer crystal. The sequential change in the exchange interactions determined by a mean-field (MF) calculation can be essentially understood by the Ruderman-Kittel-Kasuya-Yosida interaction via anisotropic Fermi surfaces, whereas the orientation of the magnetic moments will be determined by the dipole-dipole interaction. On the other hand, the MF theory predicts a much smaller critical field ${B}_{\mathrm{c}}$ than the experimentally obtained one. To discuss the deviation of ${B}_{\mathrm{c}}$ from the MF calculation, we show a possibility of a fluctuation-induced first-order transition.

GANalyze: Toward visual definitions of cognitive image properties

We introduce a framework that uses Generative Adversarial Networks (GANs) to study cognitive image properties (e.g., memorability, aesthetics, valence). Often, we do not have concrete visual definitions of what these properties entail. Starting from input noise, GANs generate a manifold of natural-looking images with fine-grained differences in their visual attributes. By navigating this manifold, we can visualize what it looks like for a particular GAN-image to become more (less) memorable. Specifically, we trained a Transformer module to learn along which direction to move a BigGAN-image’s corresponding noise vector in order to increase (or decrease) its memorability. Memorability was assessed by an off-the-shelf Assessor (MemNet). After training, we generated a test set of 1.5K “seed images”, each with four “clone images”: two modified to be more memorable (one and two “steps” forward along the learned direction) and two to be less memorable (one and two steps backward; examples in Supplemental). The assessed memorability significantly increased when stepping along the learned direction (β = 0.68, p < 0.001), suggesting training was successful. Through a behavioral repeat-detection memory experiment, we verified that our method’s manipulations indeed causally affect human memory performance (β = 1.92, p < 0.001). The seeds and their clones (i.e., "visual definitions") surfaced candidate image properties (e.g., “object size”, “colorfulness”) that may underlie memorability and were previously overlooked. These candidates correlated with the learned memorability direction. We furthermore demonstrate that stepping along a learned “object size” direction indeed increases human memorability, though less strongly (β = 0.11, p < 0.001). This showcases how the individual, causal effects of a candidate can be studied further using the same framework. Finally, we find that by substituting the Assessor, our framework can also provide visual definitions for aesthetics (β = 0.72, p < 0.001) and emotional valence (β = 0.44, p < 0.001).

Theory of the thickness dependence of the charge density wave transition in 1 T-TiTe2

Most metallic transition metal dichalcogenides undergo charge density wave (CDW) instabilities with similar or identical ordering vectors in bulk and in single layer, albeit with different critical temperatures. Metallic 1 T-TiTe2 is a remarkable exception as it shows no evidence of charge density wave formation in bulk, but it displays a stable 2 × 2 reconstruction in single-layer form. The mechanism for this 3D-2D crossover of the transition is still unclear, although strain from the substrate and the exchange interaction have been pointed out as possible formation mechanisms. Here, by performing non-perturbative anharmonic calculations with gradient corrected and hybrid functionals, we explain the thickness behaviour of the transition in 1 T-TiTe. We demonstrate that the CDW in single-layer TiTe2 occurs from the interplay of non-perturbative anharmonicity and an exchange enhancement of the electron-phonon interaction, larger in the single layer than in the bulk. Finally, we study the electronic and structural properties of the single-layer CDW phase and provide a complete description of its electronic structure, phonon dispersion as well as infrared and Raman active phonon modes.

Impact of magnetism on screw dislocations in body-centered cubic chromium

The influence of magnetism on the properties of screw dislocations in body-centered cubic chromium is investigated by means of ab initio calculations. Screw dislocations having Burgers vectors 1/2 ⟨111⟩ and ⟨100⟩ are considered, following experimental observations showing activity for both slip systems. At low temperature, chromium has a magnetic order close to antiferromagnetism along ⟨100⟩ directions, for which 1/2 ⟨111⟩ is not a periodicity vector. Hence, dislocations with Burgers vectors 1/2 ⟨111⟩ generate magnetic faults when shearing the crystal, which constrain them to coexist and move pairwise, leading to dissociated ⟨111⟩ super-dislocations. On the other side, ⟨100⟩ is a periodicity vector of the magnetic order of chromium, and no such magnetic fault are generated when ⟨100⟩ dislocations glide. Dislocation properties are computed in the magnetically ordered and non magnetic phases of chromium for comparison purposes. We report a marginal impact of magnetism on the structural properties and energies of dislocations for both slip systems. The Peierls energy barrier opposing dislocation glide in {110} planes is comparable for both 1/2 ⟨111⟩ {110} and ⟨100⟩ {110} slip systems, with lower Peierls stresses in the magnetically ordered phase of chromium.

Is Domain Knowledge Necessary for Machine Learning Materials Properties?

New featurization schemes for describing materials as composition vectors in order to predict their properties using machine learning are common in the field of Materials Informatics. However, little is known about the comparative efficacy of these methods. This work sets out to make clear which featurization methods should be used across various circumstances. Our findings include, surprisingly, that simple fractional and random-noise representations of elements can be as effective as traditional and new descriptors when using large amounts of data. However, in the absence of large datasets or for data that is not fully representative, we show that the integration of domain knowledge offers advantages in predictive ability.

Enhanced Electron-Phonon Coupling for Charge-Density-Wave Formation in La_{1.8-x}Eu_{0.2}Sr_{x}CuO_{4+δ}.

Charge density wave (CDW) correlations are prevalent in all copper-oxide superconductors. While CDWs in conventional metals are driven by coupling between lattice vibrations and electrons, the role of the electron-phonon coupling (EPC) in cuprate CDWs is strongly debated. Using Cu L_{3} edge resonant inelastic x-ray scattering, we study the CDW and Cu-O bond-stretching phonons in the stripe-ordered cuprate La_{1.8-x}Eu_{0.2}Sr_{x}CuO_{4+δ}. We investigate the interplay between charge order and EPC as a function of doping and temperature and find that the EPC is enhanced in a narrow momentum region around the CDW ordering vector. By detuning the incident photon energy from the absorption resonance, we extract an EPC matrix element at the CDW ordering vector of M≃0.36 eV, which decreases to M≃0.30 eV at high temperature in the absence of the CDW. Our results suggest a feedback mechanism in which the CDW enhances the EPC which, in turn, further stabilizes the CDW.

AUTOMATIC DETECTION AND RECOGNITION OF ROAD INTERSECTIONS FOR ROAD EXTRACTION FROM IMAGERY

Abstract. Automatic road extraction from remote sensing imagery is very useful for many applications involved with geographic information. For road extraction of urban areas, road intersections offer stable and reliable information for extraction of road network, with higher completeness and accuracy. In this paper, a segmentation-shape analysis based method is proposed to detect road intersections and their branch directions from an image. In the region of interest, it uses the contour shape of the segmented-intersection area to form a feature vector representing its geometric information. The extracted feature vector is then matched with some template vectors in order to find the best matched intersection pattern, obtain the type of intersection and the direction of connected roads. The experimental analysis are carried out with ISPRS Vaihingen and Toronto images. The experimental results show that the proposed method can extract most of the road intersections correctly. For the Vaihingen image, the the completeness and correctness are 81% and 87%, respectfully, while for the Toronto image, the the completeness and correctness are 78% and 85%, respectfully. It can help to build more correct and complete road network.

Multi-Innovation Stochastic Gradient Parameter and State Estimation Algorithm for Dual-Rate State-Space Systems with d-Step Time Delay

This paper presents a multi-innovation stochastic gradient parameter estimation algorithm for dual-rate sampled state-space systems with d-step time delay by the multi-innovation identification theory. Considering the stochastic disturbance in industrial process and using the gradient search, a multi-innovation stochastic gradient algorithm is proposed through expanding the scalar innovation into an innovation vector in order to obtain more accurate parameter estimates. The difficulty of identification is that the information vector in the identification model contains the unknown states. The proposed algorithm uses the state estimates of the observer instead of the state variables to realize the parameter estimation. The simulation results indicate that the proposed algorithm works well.

Study on the Impedance Change of Rectangular Coil Perpendicular to Conductive Cylinder

Cover meter is widely used to monitor the reinforced structures, and the response of the cover meter probe coil is used to locate the rebar, evaluate the diameter and cover depth of the rebar embedded in concrete. In order to analyze and optimize the performance of the cover meter, the analytical solution to the response of the probe is of great importance. Based on the theory of the second order vector potential (SOVP), the integral solutions to the coil coefficient and the impedance change of the rectangular coil are achieved when the normal direction of the coil is perpendicular to the axis of the conductive cylinder. The theoretical and experimental results are in excellent agreement induced not only by nonmagnetic materials but also by magnetic materials. The impedance changes of the rectangular coil and the circular coil with same sectional area are compared theoretically, as a consequence, the rectangular coil can help to find the direction of the rebar embedded in concrete. The allowed frequency of the coil is also discussed. These integral solutions to the impedance change of the rectangular coil can help the cover meter designers to optimize the coil parameters and the exciting frequency conveniently.

Wave propagation in immersed waveguide with radial inhomogeneity

In this paper, we investigate wave propagation in an inhomogeneous cylindrical waveguide immersed in a medium. To model external medium influence, we use the impedance boundary conditions on the waveguide outer wall. Some structural properties of the dispersion set are studied. To treat the problem of the waveguide forced oscillations, we solve boundary value problems for the first order vector differential equation sequentially in the Fourier transform space and then build the actual space on the basis of the residue theory. The residues for the numerically given function with the first order poles were found by considering auxiliary boundary value problems. The wave fields on the waveguide internal wall are obtained for different inhomogeneity laws and boundary conditions.

Solar sail optimal control with solar irradiance fluctuations

The optimization of a solar sail-based orbital transfer amounts to searching for the control law that minimizes the flight time. In this context, the optimal trajectory is usually determined assuming constant solar properties. However, the total solar irradiance undergoes both long-term (solar cycles) and short-term variations, and recent analyses have shown that this may have an impact on solar sailing for missions requiring an accurate thrust modulation. In this regard, the paper discusses a strategy to overcome such an issue by suitably adjusting the thrust vector in order to track a reference, optimal, transfer trajectory. In particular, the sail propulsive acceleration magnitude is modified by means of a set of electrochromic material panels, which change their optical properties on application of a suitable electric voltage. The proposed control law is validated with a set of numerical simulations that involve a classical Earth-Mars, orbit-to-orbit, heliocentric transfer.

High-field depinned phase and planar Hall effect in the skyrmion host Gd2PdSi3

For the skyrmion-hosting intermetallic Gd$_2$PdSi$_3$ with centrosymmetric hexagonal lattice and triangular net of rare earth sites, we report a thorough investigation of the magnetic phase diagram. Our work reveals a new magnetic phase with isotropic value of the critical field for all orientations, where the magnetic ordering vector $\mathbf{q}$ is depinned from its preferred directions in the basal plane. This is in contrast to the highly anisotropic behavior of the low field phases, such as the skyrmion lattice (SkL), which are easily destroyed by in-plane magnetic field. The bulk nature of the SkL and of other magnetic phases was evidenced by specific-heat measurements. Resistivity anisotropy, likely originating from partial gapping of the density of states along $\mathbf{q}$ in this RKKY magnet, is picked up via the planar Hall effect (PHE). The PHE confirms the single-$\mathbf{q}$ nature of the magnetic order when the field is in the hexagonal plane, and allows to detect the preferred directions of $\mathbf{q}$. For field aligned perpendicular to the basal plane, several scenarios for the depinned phase (DP), such as tilted conical order, are discussed on the basis of the data.

Analytic solutions for calculating the surface inclination of isotropic media and bare substrates by using reflection-based generalized ellipsometry

Using ellipsometry for curved-surface characterization requires the knowledge of the surface normal vector in order to determine material-related surface parameters like refractive index, layer thickness, or birefringence of the surface material at the incidence point, because the recorded signal depends on both the (unknown) surface normal vector n→ and material-related surface characteristics. It is convenient in ellipsometry to parametrize the surface normal vector by the angle of incidence θ and the azimuthal rotation angle ϕ. Depending on the design of the ellipsometer, there may be two angles, ϕ1 and ϕ2, necessary, which describe the azimuthal rotation before and after the light is reflected off the sample, respectively. The authors present analytic formulas to determine ϕ1 and ϕ2 for optically isotropic samples using generalized ellipsometry. The resulting measurement uncertainty is lower than that of previously known methods. Furthermore, the authors provide an analytic formula to calculate θ from the ellipsometric angles Ψ and Δ for bare substrates with known refractive index N1=n1−ik1. The formulas have been evaluated with experimental data acquired with a conventional and an imaging retroreflection-based return-path ellipsometer.