Autocorrelation Function Analysis Research Articles

Single Channel Speech Separation with Frame-based Summary Autocorrelation Function Analysis

Single channel speech separation system with frame based pitch range estimation is presented. The system decomposes the input mixture speech into frames and pitch of the speech is estimated in each frame of modulation spectrum by using summary autocorrelation function (SACF) analysis. In this method, each frame is divided into two, above and below 1 KHz and compute generalized autocorrelation for periodicity detection. By using this pitch range, a mask is created to filter the target speech from noisy mixture speech. Performance evaluation of the proposed system shows a better response compared to the existing methods.

Биофизические модели патологического и постурального тремора

Real chaotic and stochastic analysis of the two movements (tremor was considered as involuntary
 movements and tapping - arbitrary) shows them as chaotic movements (involuntary on the results of the test,
 not by the presence of the target). The authors introduce new criteria for separating these two types of motion in
 the form of matrices of pairwise comparisons of samples tremorograms and tappingrams. Identifying differences
 between the concrete (obtained continuously, during the sequential measurement) that are compared in pairs in
 one subject groups was performed using the Wilcoxon test. The increase in the number of "common" pairs of
 samples of tappingrams compared to tremorograms demonstrates a partial increase of phase-matching due to the
 afferentation and engaging mental activity. This indicates the beginning of a shift from the chaotic regime to
 stochastic. The increase in common pairs of tapping may be possible due to the change in patterns of fluctuations.
 The authors propose a new calculation of quasi-attractors of these two types of movements that allow the
 identification of the differences in the physiological state of the subject. The concrete examples of the changes in
 the parameters of the matrices of paired comparisons and quasi-attractors are demonstrated. The authors present
 a method of analysis of autocorrelation functions when partitioning the interval (-1; 1) into four parts. Using the
 analysis of the density autocorrelation functions and tremorograms and tappingrams shows a significant difference
 between involuntary movements (tremor) and arbitrary movement (tapping).

Using network dynamic fMRI for detection of epileptogenic foci.

BackgroundEpilepsy is one of the most prevalent neurological disorders. It remains medically intractable for about one-third of patients with focal epilepsy, for whom precise localization of the epileptogenic zone responsible for seizure initiation may be critical for successful surgery. Existing fMRI literature points to widespread network disturbances in functional connectivity. Per previous scalp and intracranial EEG studies and consistent with excessive local synchronization during interictal discharges, we hypothesized that, relative to same regions in healthy controls, epileptogenic foci would exhibit less chaotic dynamics, identifiable via entropic analyses of resting state fMRI time series.MethodsIn order to first validate this hypothesis on a cohort of patients with known ground truth, here we test individuals with well-defined epileptogenic foci (left mesial temporal lobe epilepsy). We analyzed voxel-wise resting-state fMRI time-series using the autocorrelation function (ACF), an entropic measure of regulation and feedback, and performed follow-up seed-to-voxel functional connectivity analysis. Disruptions in connectivity of the region exhibiting abnormal dynamics were examined in relation to duration of epilepsy and patients’ cognitive performance using a delayed verbal memory recall task.ResultsACF analysis revealed constrained (less chaotic) functional dynamics in left temporal lobe epilepsy patients, primarily localized to ipsilateral temporal pole, proximal to presumed focal points. Autocorrelation decay rates differentiated, with 100 % accuracy, between patients and healthy controls on a subject-by-subject basis within a leave-one-subject out classification framework. Regions identified via ACF analysis formed a less efficient network in patients, as compared to controls. Constrained dynamics were linked with locally increased and long-range decreased connectivity that, in turn, correlated significantly with impaired memory (local left temporal connectivity) and epilepsy duration (left temporal – posterior cingulate cortex connectivity).ConclusionsOur current results suggest that data driven functional MRI methods that target network dynamics hold promise in providing clinically valuable tools for identification of epileptic regions.

Autocorrelation optical coherence tomography for glucose quantification in blood

We report a new method for glucose monitoring in blood tissue based on the autocorrelation function (ACF) analysis in Fourier domain optical coherence tomography (FD-OCT). We have determined the changes in OCT monitoring signals’ depth to characterize the modulations in ACFs for quantitative measurements of glucose concentrations in blood. We found that an increase in the concentration of glucose in blood results in decreased OCT monitoring signal due to the increase in the refractive index of the media.

Analysis of autocorrelation and ambiguity function of complicated radiolocation signal with binary phase manipulation by Kasami sequence

Introduction. Were reviewed publications dedicated to pseudorandom sequences in radiolocation, and was made conclusion about necessity of researching in this subject. Modeling of probing radiolocation signal with binary phase manipulation by Kasami sequence. Were considered the method of forming Kasami sequences and measuring their properties. Example of generator for sequence with length 63 symbols is also shown. Technique of research. Were shown graphical results for modeling of Kasami sequence with lengths of 255 binary values. Amplitude of side maximums of autocorrelation function are measured. Modeling of ambiguity function. In this article, graphical results for modeling Kasami sequence ambiguity function are shown. These includes 3-dimensional plot of function body and its cross-sections by plane of equal ranges and planes of function level, equal to 0.9, 0.7, 0.5, 0.4, 0.3, 0.2 and 0.1 of its maximum. Were analyzed amplitude of side maximums, cross-section form and relative values of range and speed resolution. Conclusion. These articles contain analyses of characteristics of previously shown plots. These analyses include estimation of total value of side maximums of ambiguity function of Kasami sequence, ambiguity function cross-section form analysis and conclusion about ability of using these sequences as probe signal in radiolocation.

Расчет нейтральных меридиональных ветров в средних широтах с помощью Иркутского радара НР

The paper sequentially presents technique for determining velocity of meridional neutral winds from the Irkutsk Incoherent Scatter Radar (IISR) data. Due to IISR specific features effective at other IS radars, techniques for determining ionosphere parameters, in particular plasma drift velocities, resulted in considerable variance of defined parameters. To measure the plasma drift velocity taking into account such IISR features, we have developed a special technique based on phase analysis of autocorrelation function of incoherent scatter signal. The technique needs to be tested, and for this purpose, an experiment was carried out to measure velocities of low-orbit satellites. However, methods for meridional neutral wind calculations used by many authors [Evans, 1970] with the use of drift velocities obtained before, resulted in great disagreements with empirical HWM93 and HWM07 wind models. In addition, simultaneous measurements at two frequencies at IISR showed that it was difficult to explain such differences without taking into account the cross-field movements. Possible underestimation of the impact of movements generated by electric fields can result in serious error in determining wind velocities. The paper considers improvements for methods of calculating winds, and shows that the results obtained with it are in a better agreement with wind models.

The features of long-term variability of water thermohaline characteristics in the North Atlantic and Greenland Sea in 1950-2012

Considered are the results of studying the variability of water thermohaline structure at different time scales. Time series and cross-sections are analyzed in the chosen test areas and trends are obtained. The analysis of autocorrelation functions and spectral density of the data enabled identifying the main periods of the variability of temperature and salinity in each area under study. The area of the “dome” of bottom and deep waters is considered in detail. The TS-analysis and analysis of spatial cross-sections were carried out to study comprehensively the processes of salinization and warming of water in the recent two decades. The effects of advection from the adjacent areas are also studied by means of cross-correlation and wavelet analyses.

Local structure analysis in ab initio liquid water

Within the framework of density functional theory, the inclusion of exact exchange and non-local van der Waals/dispersion (vdW) interactions is crucial for predicting a microscopic structure of ambient liquid water that quantitatively agrees with experiment. In this work, we have used the local structure index (LSI) order parameter to analyse the local structure in such highly accurate ab initio liquid water. At ambient conditions, the LSI probability distribution, P(I ), was unimodal with most water molecules characterised by more disordered high-density-like local environments. With thermal excitations removed, the resultant bimodal P(I ) in the inherent potential energy surface (IPES) exhibited a 3:1 ratio between high-density- and low-density-like molecules, with the latter forming small connected clusters amid the predominant population. By considering the spatial correlations and hydrogen bond network topologies among water molecules with the same LSI identities, we demonstrate that the signatures of the experimentally observed low- and high-density amorphous phases of ice are present in the IPES of ambient liquid water. Analysis of the LSI autocorrelation function uncovered a persistence time of ∼ 4 ps – a finding consistent with the fact that natural thermal fluctuations are responsible for transitions between these distinct yet transient local aqueous environments in ambient liquid water.

Structural Properties and Dynamics of Thiophene in Sub/Supercritical Carbon Dioxide from Car-Parrinello Molecular Dynamics Simulations.

Structrual and dynamic properties of thiophene (C4H4S) in supercritical carbon dioxide were studied using Car-Parrinello molecular dynamics simulations. The geometries and energies optimized for the thiophene-CO2 complex show a stable C-H···O hydrogen bond interactions both in gas phase and in supercritical CO2. The radial distribution functions of CO2 around thiophene in the supercritical phase state show a correlation suggesting C-H···O hydrogen bond and S···C interaction. Local structural properties of the mixtures were investigated by angular-radial distributions and spatial distribution functions. The results show a mutually parallel arrangement between the thiophene plane and CO2 molecules at short distances and a high probability of the thiophene being located in the radial directions of the CO2 molecules. The decay of orientational correlations at 318.15 K shows slower relaxation compared to those of 298.15 K for first and second rank correlations. The vibrations of CO2 and thiophene molecules have been examined through an analysis of the velocity autocorrelation functions of the atoms. The C-H stretching modes of thiophene in the isolated configuration are less red-shifted and have a much narrower frequency range than that in the mixtures.

Mixing dynamics for easy flowing powders in a lab scale Turbula® mixer

Powder mixing is one of the most common unit operations in industries handling powders. However, due to a lack of scientific knowledge on the behavior of powders inside a mixer, the optimization of a powder mixing step typically involves an experimental lab work in order to be transposed to a larger scale. To simplify this huge work, improving our understanding of the mixing dynamics remains some of the main industrial issues.In this work, the mixing dynamics of an easy flowing powder made of lactose and couscous have been studied in the Turbula® mixer T2F. According to process parameters, we focused on the definition of mixing modes and we sought to identify the main mixing and segregation mechanisms at play for each mode, based on the analysis of mixing kinetics and autocorrelation functions. This work is a first step toward the definition of scale-up laws. The ultimate aim is to build predictive models to determine process parameters for obtaining the desired quality of mixing according to powder characteristics, flow properties and industrial constraints.

Characteristics of extreme heavy precipitation events occurring in the area of Cracow (Poland)

The variability of extremely heavy precipitation events with duration of 120 min occurring in the area of Cracow, southern Poland was assessed. The analysis was performed using time series of maximum annual precipitation events with durations t = 5, 10, 15, 30, 60, and 120 min, recorded at the Botanical Garden station at the Jagiellonian University in the period of 1906–1990. The periodicity of precipitation was analyzed using the autocorrelation function and Fourier spectral density analysis. The Probable Maximum Precipitation (PMP) was calculated by Hershfield’s statistical method. The analysis of the autocorrelation function of sequences and the Fourier spectral density revealed a clear periodicity of the maximum precipitation. For precipitation with t = 60 min, the maximum values occur every 9 years, but also shorter periods (3-year) may be observed. The PMP values calculated for Cracow differ significantly from the values calculated using the probability distribution, as well as from the ones observed and they increase with increasing precipitation duration. The differences between the PMP and probable as well as observed precipitation tend to decrease with increasing duration of precipitation.

The Selection of the Best Estimation Model of the Composite Index on Stock Exchanges in Five ASEAN Countries Using Box Jenkins Method

This paper aims to obtain the Estimation Model of the Composite Index on Stock Exchanges in five countries of ASEAN. Those countries are Indonesia, Malaysia, Singapore, The Philippines, and Thailand. The estimation method used in this paper is Box Jenkins Method. The used data in building this estimation model are weekly data of composite index from 2000 to 2012 obtained from Blomberg. The best model is expected to be usable as one of references in estimating the composite index on those five stock exchanges. The best model selection is respectively begun by looking at the stationary data. Differencing will be conducted if the data are not stationary until they are. The next step is to do Autocorrelation Function (ACF) and Partial Autocorrelation Function (PAFC) analysis to gain ARIMA model. The best model selection is conducted by observing the coefficient significance on each model and select model with the smallest sum square error (SSE). Then, to ensure the selected model meets the assumption of white noise and normally distributed, the model diagnosis and normality test will be done using Kolmogorof Smirnof. Based on the analysis above, it can be decided that the estimation model of the best composite index from the five stock exchanges is ARIMA (2,1,1) for Indonesia, ARIMA (6,1,2) for Malaysia, ARIMA (5,1,5) for Singapore, ARIMA (2,1,0) for The Philippines, and ARIMA (4,1,0) for Thailand.

Spherical silicon optical resonators: Possible applications to biosensing

We observed whispering gallery modes in the 90∘ elastic light scattering and 0∘ transmission spectra of a 500 μm silicon microsphere in the near-infrared telecommunication wavelengths. The whispering gallery modes have quality factors on the order of 105. An autocorrelation function analysis of the 90∘ elastic light scattering and 0∘ transmission spectra is performed. The differential autocorrelation of the 90∘ elastic light scattering and 0∘ transmission spectra reveal a spectral mode spacing 0.23 nm of the whispering gallery modes, which can be used for spectral calibration as well as the calibration of the microsphere size. The spectral shift of the whispering gallery modes can be measured for biosensing applications. An estimation analysis is performed for the adsorption of 30 DNA base pairs on the microsphere results in a wavelength shift of 0.57 nm, which is approximately 40 times the linewidth of the whispering gallery modes. This high sensitivity heralds silicon microspheres as possible candidates for biosensing applications.

PLL bandwidth and noise in 100 Hz GPS measurements

The impact of the phase-locked loop (PLL) bandwidth on the noise and correlations of GPS measurements sampled with a 100 Hz rate has been investigated using short and long baselines, and stationary or moving GPS rovers recording oscillations with known characteristics relative to "true" reference values. Data have been collected under various satellite constellations using various values of PLL bandwidth, particularly 25, 50, 100 Hz, and were processed in differential mode using different software packages. Analysis of standard deviations, spectra and autocorrelation functions of the differences between recorded and true displacements revealed that an increase in the PLL bandwidth leads to reduction in correlations and increase in low- and high-frequency noise of 100 Hz data. Optimal results can be obtained using either a pre-set 50 Hz PLL bandwidth or a 100 Hz PLL bandwidth combined with a posteriori band-pass filtering of the coordinates. Such optimal results permit accurate recording of high-frequency (>5 Hz), dynamic displacements of a few millimeters and indicate that 100 Hz data are useful for monitoring high-frequency structural vibrations, and also strong earthquake motions and high-frequency movements of vehicles.

Spin-Current Autocorrelations from Single Pure-State Propagation

We demonstrate that the concept of quantum typicality allows for significant progress in the study of real-time spin dynamics and transport in quantum magnets. To this end, we present a numerical analysis of the spin-current autocorrelation function of the antiferromagnetic and anisotropic spin-1/2 Heisenberg chain as inferred from propagating only a single pure state, randomly chosen as a "typical" representative of the statistical ensemble. Comparing with existing time-dependent density-matrix renormalization group data, we show that typicality is fulfilled extremely well, consistent with an error of our approach, which is perfectly under control and vanishes in the thermodynamic limit. In the long-time limit, our results provide for a new benchmark for the enigmatic spin Drude weight, which we obtain from chains as long as L=33 sites, i.e., from Hilbert spaces of dimensions almost O(104) larger than in existing exact-diagonalization studies.

Forecasting Rail Transport Petroleum Consumption Using an Integrated Model of Autocorrelation Functions-Artificial Neural Network

This paper presents the application of time-series and artificial neural network for improvement of energy forecasting in rail transport section. An integrated artificial neural network (ANN) model is presented that uses autocorrelation and partial autocorrelation functions to determine the best input variables for ANN. The proposed ANN uses autocorrelation function (ACF) and partial autocorrelation function (PACF) extracted from time-series data to select appropriate inputs for ANN. For validation of the ANN results, they are compared with two auto regressive models using analysis of variance (ANOVA) technique. Weekly total gasoil consumption in Iran railway transportation from January 2009 to October 2011 is used for constructing and comparison time-series and ANN models. Result shows that ANN with inputs extracted from ACF and PACF analysis, performs well for estimation of railway gasoil consumption. The interaction between input variables as well as non-linearity in consumption data is properly handled by the proposed integrated ANN.

The effect of cooling rate on the wear performance of a ZrCuAlAg bulk metallic glass

In the present work, the local atomic ordering and the wear performance of ZrCuAlAg bulk metallic glass (BMG) samples with different diameters have been studied using transmission electron microscopy (TEM) plus autocorrelation function analysis, and pin-on-disc dry sliding wear experiments. Differential scanning calorimetry and TEM studies show that smaller diameter BMG sample has higher free volume and less local atomic ordering. The wear experiments demonstrate that with the same chemical composition, the smaller BMG sample exhibits higher coefficient of friction, higher wear rate, and rougher worn surface than those of the larger ones. Compared with larger BMG sample, the faster cooling rate of the smaller sample results in looser atomic configuration with more free volume, which facilitates the formation of the shear bands, and thus leads to larger plasticity and lower wear resistance. The results provide more quantitative understanding on the relationship among the cooling rate, the local atomic ordering, and the wear performance of BMGs.

FERMI LARGE AREA TELESCOPE DETECTION OF GRAVITATIONAL LENS DELAYED γ-RAY FLARES FROM BLAZAR B0218+357

Using data from the Fermi Large Area Telescope (LAT), we report the first clear gamma-ray measurement of a delay between flares from the gravitationally lensed images of a blazar. The delay was detected in B0218+357, a known double-image lensed system, during a period of enhanced gamma-ray activity with peak fluxes consistently observed to reach >20-50 times its previous average flux. An auto-correlation function analysis identified a delay in the gamma-ray data of 11.46 +/- 0.16 days (1 sigma) that is ~1 day greater than previous radio measurements. Considering that it is beyond the capabilities of the LAT to spatially resolve the two images, we nevertheless decomposed individual sequences of superposing gamma-ray flares/delayed emissions. In three such ~8-10 day-long sequences within a ~4-month span, considering confusion due to overlapping flaring emission and flux measurement uncertainties, we found flux ratios consistent with ~1, thus systematically smaller than those from radio observations. During the first, best-defined flare, the delayed emission was detailed with a Fermi pointing, and we observed flux doubling timescales of ~3-6 hrs implying as well extremely compact gamma-ray emitting regions.

Wind-Flow Dynamics Over a Vineyard

Wind-flow dynamics has been extensively studied over horizontally uniform canopies, but agricultural plantations structured in rows such as vineyards have received less attention. Here, the wind flow over a vineyard is studied in neutral stratification from both large-eddy simulation (LES) and in situ measurements. The impact of row structure on the wind dynamics is investigated over a range of wind directions from cross-row to down-row, and a typical range of row aspect ratio (row separation/height ratio). It is shown that the mean flow over a vineyard is similar to that observed in uniform canopies, especially for wind directions from cross-row to diagonal. For down-row winds, the mean flow exhibits noticeable spatial variability across each elementary row-gap pattern, as the wind is channeled in the inter-row. This spatial variability increases with the aspect ratio. With down-row winds the turbulent structures are also more intermittent and generate larger turbulent kinetic energy and momentum flux. The displacement height and roughness length of the vineyard vary with the aspect ratio in a way similar to their variation with canopy density in uniform canopies. Both parameters take smaller values in down-row wind flow, for which the canopy appears more open. The analysis of velocity spectra and autocorrelation functions shows that vineyard canopies share similar features to uniform canopies in terms of turbulent coherent structures, with only minor changes with wind direction.

Sub-THz specific relaxation times of hydrogen bond oscillations in E.coli thioredoxin. Molecular dynamics and statistical analysis.

Hydrogen bonds (H-bonds) in biological macromolecules are important for the molecular structure and functions. Since interactions via hydrogen bonds are weaker than covalent bonds, it can be expected that atomic movements involving H-bonds have low frequency vibrational modes. Sub-Terahertz (sub-THz) vibrational spectroscopy that combines measurements with molecular dynamics (MD) computational prediction has been demonstrated as a promising approach for biological molecule characterization. Multiple resonance absorption lines have been reported. The knowledge of relaxation times of atomic oscillations is critical for the successful application of THz spectroscopy for hydrogen bond characterization. The purpose of this work is to use atomic oscillations in the 0.35-0.7 THz range, found from molecular dynamic (MD) simulations of E.coli thioredoxin (2TRX), to study relaxation dynamics of two intra-molecular H-bonds, OH-N and OH-C. Two different complimentary techniques are used in this study, one is the analysis of the statistical distribution of relaxation time and dissipation factor values relevant to low frequency oscillations, and the second is the analysis of the autocorrelation function of low frequency quasi-periodic movements. By studying hydrogen bond atomic displacements, it was found that the atoms are involved in a number of collective oscillations, which are characterized by different relaxation time scales ranging from 2-3 ps to more than 150 ps. The existence of long lasting relaxation processes opens the possibility to directly observe and study H-bond vibrational modes in sub-THz absorption spectra of bio-molecules if measured with an appropriate spectral resolution. The results of measurements using a recently developed frequency domain spectroscopic sensor with a spectral resolution of 1 GHz confirm the MD analysis.