Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “blind source separation”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

36 records · Page 2

Complex Signal Kurtosis and Independent Component Analysis for Wideband Radio Frequency Interference Detection

Radio-frequency interference (RFI) has negatively implicated scientific measurements across a wide variation passive remote sensing satellites. This has been observed in the L-band radiometers SMOS, Aquarius and more recently, SMAP [1, 2]. RFI has also been observed at higher frequencies such as K band [3]. Improvements in technology have allowed wider bandwidth digital back ends for passive microwave radiometry. A complex signal kurtosis radio frequency interference detector was developed to help identify corrupted measurements [4]. This work explores the use of ICA (Independent Component Analysis) as a blind source separation technique to pre-process radiometric signals for use with the previously developed real and complex signal kurtosis detectors.

microwave↗

RFI Detection and Mitigation using Independent Component Analysis as a Pre-Processor

Radio-frequency interference (RFI) has negatively impacted scientific measurements of passive remote sensing satellites. This has been observed in the L-band radiometers Soil Moisture and Ocean Salinity (SMOS), Aquarius and more recently, Soil Moisture Active Passive (SMAP). RFI has also been observed at higher frequencies such as K band. Improvements in technology have allowed wider bandwidth digital back ends for passive microwave radiometry. A complex signal kurtosis radio frequency interference detector was developed to help identify corrupted measurements. This work explores the use of Independent Component Analysis (ICA) as a blind source separation (BSS) technique to pre-process radiometric signals for use with the previously developed real and complex signal kurtosis detectors.

independent component analysis↗

An RFI Detection Algorithm for Microwave Radiometers Using Sparse Component Analysis

Radio Frequency Interference (RFI) is a threat to passive microwave measurements and if undetected, can corrupt science retrievals. The sparse component analysis (SCA) for blind source separation has been investigated to detect RFI in microwave radiometer data. Various techniques using SCA have been simulated to determine detection performance with continuous wave (CW) RFI.

radiometers↗

An RFI Detection Algorithm for Microwave Radiometers Using Sparse Component Analysis

Radio Frequency Interference (RFI) is a threat to passive microwave measurements and if undetected, can corrupt science retrievals. The sparse component analysis (SCA) for blind source separation has been investigated to detect RFI in microwave radiometer data. Various techniques using SCA have been simulated to determine detection performance with continuous wave (CW) RFI.

radiometers↗

Independent Component Analysis for Preprocessing Optical Signals in Support of Multi-User Communication

In an effort to construct an optical transceiver for supporting multi-point communication, researchers developed large field-of-view FSO transceivers with wide apertures for facilitating rapid acquisition and optical signal tracking. The design is constructed from fiber-bundle(s) intelligently arranged to provide multiple optical pathways within a single transceiver, with each pathway limited to a particular optical directionality. Because designs with wide apertures are susceptible to receiving several optical signals simultaneously from multiple transmitters, accurately decoding individual signals and achieving true multi-user communication can be difficult. The work detailed in this paper investigated the use of blind source separation techniques, mainly independent component analysis (ICA), to estimate transmitted signals from their observed, combined mixtures sans information. Effects of signal power, format, data rate, wavelength, and turbulence severity on signal separation, as well as signal demodulation accuracy was also analyzed using experimental optical signals and corresponding mixtures. Signal generators, laser sources, turbulence emulator, beam profiler, photodiodes, and oscilloscopes comprised the experiment setup. Optical signals with varying power, format, and wavelength were generated and propagated through turbulence with varying degree severity, and finally received by the fiber-bundle-based optical transceiver. Results demonstrate that an ICA technique can accurately process the combined signals into corresponding individual components. The investigation further identified signal characteristics (e.g., power levels, format, data rate) and turbulence severity under which ICA is unable to accurately perform signal separations.

Optics↗

System and method for automated extraction of high resolution structural dynamics from video

A method for extracting vibrational modes of a structure includes: receiving a plurality of video frames, each of the video frames including a plurality of pixels; decomposing each of the video frames on a plurality of spatial scales in accordance with complex steerable pyramid filters to obtain a filter response for each of the spatial scales; computing a plurality of local phases of the pixels of each frame; removing a temporal mean from each frame to obtain a plurality of factored vibration motion functions; performing principal component analysis on the factored vibration motion functions to obtain principal components; blind source separating the principal components to compute a plurality of modal coordinates; computing frequency and damping ratios in accordance with the modal coordinates; and outputting the computed frequency and damping ratios.

42 ENGINEERING↗

Source identification by non-negative matrix factorization combined with semi-supervised clustering

Machine-learning methods and apparatus are provided to solve blind source separation problems with an unknown number of sources and having a signal propagation model with features such as wave-like propagation, medium-dependent velocity, attenuation, diffusion, and/or advection, between sources and sensors. In exemplary embodiments, multiple trials of non-negative matrix factorization are performed for a fixed number of sources, with selection criteria applied to determine successful trials. A semi-supervised clustering procedure is applied to trial results, and the clustering results are evaluated for robustness using measures for reconstruction quality and cluster separation. The number of sources is determined by comparing these measures for different trial numbers of sources. Source locations and parameters of the signal propagation model can also be determined. Disclosed methods are applicable to a wide range of spatial problems including chemical dispersal, pressure transients, and electromagnetic signals, and also to non-spatial problems such as cancer mutation.

97 MATHEMATICS AND COMPUTING↗

Source identification by non-negative matrix factorization combined with semi-supervised clustering

Machine-learning methods and apparatus are provided to solve blind source separation problems with an unknown number of sources and having a signal propagation model with features such as wave-like propagation, medium-dependent velocity, attenuation, diffusion, and/or advection, between sources and sensors. In exemplary embodiments, multiple trials of non-negative matrix factorization are performed for a fixed number of sources, with selection criteria applied to determine successful trials. A semi-supervised clustering procedure is applied to trial results, and the clustering results are evaluated for robustness using measures for reconstruction quality and cluster separation. The number of sources is determined by comparing these measures for different trial numbers of sources. Source locations and parameters of the signal propagation model can also be determined. Disclosed methods are applicable to a wide range of spatial problems including chemical dispersal, pressure transients, and electromagnetic signals, and also to non-spatial problems such as cancer mutation.

Alexandrov, Boian S.↗

A Contextually Supervised Optimization-Based HVAC Load Disaggregation Methodology

This paper presents a novel contextually supervised optimization-based approach for disaggregating heating, ventilation, and air-conditioning (HVAC) loads using smart meter or Supervisory Control and Data Acquisition data. To disaggregate the load into HVAC loads, large and infrequently used loads (LIUL), and base loads, we formulate an optimization problem to minimize a set of five loss terms, consisting of the reconstruction errors of the overall load profile, the ramp rate losses, and three distinct loss functions linked with the HVAC load, base load, and LIUL, respectively. To enhance accuracy, we incorporate two forms of contextual information into the problem formulation. First, we utilize mutual information to estimate HVAC energy consumption. Second, we employ a base load dictionary to constrain HVAC load estimation errors. The obtained HVAC load profiles are fine-tuned by abnormal ramp detection followed by binary hypothesis testing. Here, the proposed method is developed and tested using sub-metered residential and commercial building data. Simulation results show that the proposed method outperforms existing methods across various data resolutions and load aggregation levels, showing excellent transferability and generalizability.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Analysis and visualization of single-trial event-related potentials

In this study, a linear decomposition technique, independent component analysis (ICA), is applied to single-trial multichannel EEG data from event-related potential (ERP) experiments. Spatial filters derived by ICA blindly separate the input data into a sum of temporally independent and spatially fixed components arising from distinct or overlapping brain or extra-brain sources. Both the data and their decomposition are displayed using a new visualization tool, the "ERP image," that can clearly characterize single-trial variations in the amplitudes and latencies of evoked responses, particularly when sorted by a relevant behavioral or physiological variable. These tools were used to analyze data from a visual selective attention experiment on 28 control subjects plus 22 neurological patients whose EEG records were heavily contaminated with blink and other eye-movement artifacts. Results show that ICA can separate artifactual, stimulus-locked, response-locked, and non-event-related background EEG activities into separate components, a taxonomy not obtained from conventional signal averaging approaches. This method allows: (1) removal of pervasive artifacts of all types from single-trial EEG records, (2) identification and segregation of stimulus- and response-locked EEG components, (3) examination of differences in single-trial responses, and (4) separation of temporally distinct but spatially overlapping EEG oscillatory activities with distinct relationships to task events. The proposed methods also allow the interaction between ERPs and the ongoing EEG to be investigated directly. We studied the between-subject component stability of ICA decomposition of single-trial EEG epochs by clustering components with similar scalp maps and activation power spectra. Components accounting for blinks, eye movements, temporal muscle activity, event-related potentials, and event-modulated alpha activities were largely replicated across subjects. Applying ICA and ERP image visualization to the analysis of sets of single trials from event-related EEG (or MEG) experiments can increase the information available from ERP (or ERF) data. Copyright 2001 Wiley-Liss, Inc.

Non-NASA Center↗

Comparative Soot Diagnostics Experiment Looks at the Smoky World of Microgravity Combustion

From an economic standpoint, soot is one of the most important combustion intermediates and products. It is a major industrial product and is the dominant medium for radiant heat transport in most flames used to generate heat and power. The nonbuoyant structure of most flames of practical interest (turbulent flames) makes the understanding of soot processes in microgravity flames important to our ability to predict fire behavior on Earth. In addition, fires in spacecraft are considered a credible possibility. To respond to this risk, NASA has flown fire (or smoke) detectors on Skylab and the space shuttles and included them in the International Space Station design. The design of these detectors, however, was based entirely on normal gravity (1g) data. The detector used in the shuttle fleet is an ionization detector, whereas the system planned for the space station uses forward scattering of near-infrared light. The ionization detector, which is similar to smoke detectors used in homes, has a comparative advantage for submicron particulates. In fact, the space shuttle model uses a separation system that makes it blind to particles larger than a micron (believed to be dust). In the larger size range, the lightscattering detector is most sensitive. Without microgravity smoke data, the difference in the particle size sensitivities of the two detectors cannot be evaluated. As part of the Comparative Soot Diagnostics (CSD) experiment, these systems were tested to determine their response to particulates generated during long periods of low gravity. This experiment provided the first such measurements toward understanding soot processes on Earth and for designing and implementing improved spacecraft smoke detection systems. The objectives of CSD were to examine how particulates form from a variety of sources and to quantify the performance of several diagnostic techniques. The sources tested included four overheated materials (paper, silicone rubber, Teflon-coated (DuPont) wire, and Kapton-coated (DuPont) wires), each tested at three heating rates, and a candle tested at three air velocities. Paper, silicone rubber, and wire insulation, materials found in spacecraft crew cabins, were selected because of their different smoke properties. The candle yielded hydrocarbon soot typical of many 1g flames. Four diagnostic techniques were employed: thermophoretic sampling collected particulates for size analysis; laser light extinction measurements near the source tallied total particulate production; and laser light scattering and ionization detector measurements far from the particulate source provided data for evaluating the performance of smoke detection systems for these particulate sources.

Urban, David L.↗

Probing parity violation with the four-point correlation function of BOSS galaxies

Parity-violating physics in the early universe can leave detectable traces in late-time observables. While vector- and tensor-type parity violation can be observed in the B -modes of the cosmic microwave background, scalar-type signatures are visible only in the four-point correlation function (4PCF) and beyond. This work presents a blind test for parity violation in the 4PCF of the baryon oscillation spectroscopic survey (BOSS) CMASS sample, considering galaxy separations in the range [20; 160]h -1 Mpc The parity-odd 4PCF contains no contributions from standard Λ CDM physics and can be efficiently measured using recently developed estimators. Data are analyzed using both a nonparametric rank test (comparing the BOSS 4PCFs to those of realistic simulations) and a compressed χ 2 analysis, with the former avoiding the assumption of a Gaussian likelihood. These find similar results, with the rank test giving a detection probability of 99.6% (2.9 σ). This provides significant evidence for parity violation, from either cosmological sources or systematics. We perform a number of systematic tests: although these do not reveal any observational artifacts, we cannot exclude the possibility that our detection is caused by the simulations not faithfully representing the statistical properties of the BOSS data. Our measurements can be used to constrain physical models of parity violation. As an example, we consider a coupling between the inflaton and a U (1) gauge field and place bounds on the latter’s energy density, which are several orders of magnitude stronger than those previously reported. Upcoming probes such as DESI and Euclid will reveal whether our detection of parity violation is due to new physics, and strengthen the bounds on a variety of models.

79 ASTRONOMY AND ASTROPHYSICS↗

Hyper Suprime-Cam Year 3 results: Cosmology from galaxy clustering and weak lensing with HSC and SDSS using the minimal bias model

Here, we present cosmological parameter constraints from a blind joint analysis of three two-point correlation functions measured from the Year 3 Hyper Suprime-Cam (HSC-Y3) imaging data, covering about 416 deg 2 , and the SDSS DR11 spectroscopic galaxies spanning the redshift range [0.15, 0.70]. We subdivide the SDSS galaxies into three luminosity-cut, and therefore nearly volume-limited samples separated in redshift, each of which acts as a large-scale structure tracer characterized by the measurement of the projected correlation function, w p ⁡(R). We also use the measurements of the galaxy-galaxy weak-lensing signal Δ⁢Σ⁡(R) for each of these SDSS samples which act as lenses for a secure sample of source galaxies selected from the HSC-Y3 shape catalog based on their photometric redshifts. We combine these measurements with the cosmic shear correlation functions, ξ ± ⁡($\vartheta$) measured for our HSC source sample. We model these observables with the minimal bias model of the galaxy clustering observables in the context of a flat Λ⁢CDM cosmology. We use conservative scale cuts, R >12 and 8⁢ℎ -1 Mpc for Δ⁢Σ and w p , respectively, where the minimal bias model is valid, in addition to conservative prior on the residual bias in the mean redshift of the HSC photometric source galaxies. We present various validation tests of our model as well as analysis methods. Our baseline analysis yields S 8 = $0.77⁢5^{+0.043}_{-0.038}$ (68% C.I.) for the Λ⁢CDM model, after marginalizing over uncertainties in other parameters. Our value of S 8 is consistent with that from the Planck 2018 data, but the credible interval of our result is still relatively large. We show that various internal consistency tests based on different splits of the data are passed. Our results are statistically consistent with those of a companion paper, which extends this analysis to smaller scales with an emulator-based halo model, using Δ⁢Σ⁡(R) and w p ⁡(R) down to R >3 and 2⁢ℎ -1 Mpc, respectively.

79 ASTRONOMY AND ASTROPHYSICS↗

A Fermi Gamma-Ray Burst Monitor Search for Electromagnetic Signals Coincident with Gravitational-Wave Candidates in Advanced LIGOs First Observing Run

We present a search for prompt gamma-ray counterparts to compact binary coalescence gravitational wave (GW)candidates from Advanced LIGO's first observing run (O1). As demonstrated by the multimessenger observations ofGW170817/GRB 170817A, electromagnetic and GW observations provide complementary information about theastrophysical source, and in the case of weaker candidates, may strengthen the case for an astrophysical origin. Here weinvestigate low-significance GWcandidates from the O1 compact binary coalescence searches using the Fermi Gamma-Ray Burst Monitor (GBM), leveraging its all sky and broad energy coverage. Candidates are ranked and compared tobackground to measure the significance. Those with false alarm rates (FARs) of less than 105 Hz (about one per day,yielding a total of 81 candidates) are used as the search sample for gamma-ray follow-up. No GW candidates werefound to be coincident with gamma-ray transients independently identified by blind searches of the GBM data. Inaddition, GW candidate event times were followed up by a separate targeted search of GBM data. Among the resultingGBM events, the two with the lowest FARs were the gamma-ray transient GW150914-GBM presented in Connaughtonet al. and a solar flare in chance coincidence with a GW candidate.

Burns, E.↗

Galileo Probe Measurements of Thermal and Solar Radiation Fluxes in the Jovian Atmosphere

The Galileo probe net flux radiometer (NFR) measured radiation fluxes in Jupiter's atmosphere from about 0.44 to 14 bars, using five spectral channels to separate solar and thermal components. Onboard calibration results confirm that the NFR responded to radiation approximately as expected. NFR channels also responded to a superimposed thermal perturbation, which can be approximately removed using blind channel measurements and physical constraints. Evidence for the expected NH3 cloud was seen in the spectral character of spin-induced modulations of the direct solar beam signals. These results are consistent with an overlying cloud of small NH3 ice particles (0.5-0.75 microns in radius) of optical depth 1.5-2 at 0.5 microns. Such a cloud would have so little effect on thermal fluxes that NFR thermal channels provide no additional constraints on its properties. However, evidence for heating near 0.45 bar in the NFR thermal channels would seem to require either an additional opacity source beyond this small-particle cloud, implying a heterogeneous cloud structure to avoid conflicts with solar modulation results, or a change in temperature lapse rate just above the probe measurements. The large thermal flux levels imply water vapor mixing ratios that are only 6% of solar at 10 bars, but possibly increasing with depth, and significantly subsaturated ammonia at pressures less than 3 bars. If deep NH3 mixing ratios at the probe entry site are 3-4 times ground-based inferences, as suggested by probe radio signal attenuation, then only half as much water is needed to match NFR observations. No evidence of a water cloud was seen near the 5-bar level. The 5-microns thermal channel detected the presumed NH4SH cloud base near 1.35 bars. Effects of this cloud were also seen in the solar channel upflux measurements but not in the solar net fluxes, implying that the cloud is a conservative scatterer of sunlight. The minor thermal signature of this cloud is compatible with particle radii near 3 gm, but it cannot rule out smaller particles. Deeper than about 3 bars, solar channels indicate unexpectedly large absorption of sunlight at wavelengths longer than 0.6 microns, which might be due to unaccounted-for absorption by NH3 between 0.65 and 1.5 microns.

Sromovsky, L. A.↗

Galileo Probe Measurements of Thermal and Solar Radiation Fluxes in the Jovian Atmosphere

The Galileo probe net flux radiometer (NFR) measured radiation fluxes in Jupiter's atmosphere from about 0.44 to 14 bars, using five spectral channels to separate solar and thermal components. Onboard calibration results confirm that the NFR responded to radiation approximately as expected. NFR channels also responded to a superimposed thermal perturbation, which can be approximately removed using blind channel measurements and physical constraints. Evidence for the expected NH3 cloud was seen in the spectral character of spin-induced modulations of the direct solar beam signals. These results are consistent with an overlying cloud of small NH3 ice particles (0.5-0.75 microns in radius) of optical depth 1.5-2 at 0.5 microns. Such a cloud would have so little effect on thermal fluxes that NFR thermal channels provide no additional constraints on its properties. However, evidence for heating near 0.45 bar in the NFR thermal channels would seem to require either an additional opacity source beyond this small-particle cloud, implying a heterogeneous-cloud structure to avoid conflicts with solar modulation results, or a change in temperature lapse rate just above the probe measurements. The large thermal flux levels imply water vapor mixing ratios that are only 6% of solar at 10 bars, but possibly increasing with depth, and significantly subsaturated ammonia at pressures less than 3 bars. If deep NH3 mixing ratios at the probe entry site are 3-4 times ground-based inferences, as suggested by probe radio signal attenuation, then only half as much water is needed to match NFR observations. No evidence of a water cloud was seen near the 5-bar level. The 5 microns thermal channel detected the presumed NH4SH cloud base near 1.35 bars. Effects of this cloud were also seen in the solar channel upflux measurements but not in the solar net fluxes, implying that the cloud is a conservative scatterer of sunlight. The minor thermal signature of this cloud is compatible with particle radii near 3 microns, but it cannot rule out smaller particles. Deeper than about 3 bars, solar channels indicate unexpectedly large absorption of sunlight at wavelengths longer than 0.6 microns, which might be due to unaccounted-for absorption by NH3 between 0.65 and 1.5 microns.

Sromovsky, L. A.↗

Signal-preserving CMB component separation with machine learning

Analysis of microwave sky signals, such as the cosmic microwave background, often requires component separation using multifrequency methods, whereby different signals are isolated according to their different frequency behaviors. Many so-called blind methods, such as the internal linear combination (ILC), make minimal assumptions about the spatial distribution of the signal or contaminants, and only assume knowledge of the frequency dependence of the signal. The ILC produces a minimum-variance linear combination of the measured frequency maps. In the case of Gaussian, statistically isotropic fields, this is the optimal linear combination, as the variance is the only statistic of interest. However, in many cases the signal we wish to isolate, or the foregrounds we wish to remove, are non-Gaussian and/or statistically anisotropic (in particular for the case of Galactic foregrounds). In such cases, it is possible that machine learning (ML) techniques can be used to exploit the non-Gaussian features of the foregrounds and thereby improve component separation. However, many ML techniques require the use of complex, difficult-to-interpret operations on the data. We propose a hybrid method whereby we train an ML model using only combinations of the data that , and combine the resulting ML-predicted foreground estimate with the ILC solution to reduce the error from the ILC. We demonstrate our methods on simulations of extragalactic temperature and Galactic polarization foregrounds and show that our ML model can exploit non-Gaussian features, such as point sources and spatially varying spectral indices, to produce lower-variance maps than ILC—e.g., reducing the variance of the B-mode residual by factors of up to 5—while preserving the signal of interest in an unbiased manner. Moreover, we often find improved performance even when applying our ML technique to foreground models on which it was not trained. Published by the American Physical Society 2025

McCarthy, Fiona (ORCID:0000000253893565)↗

Resolving local structural motifs across the phase evolution of zinc titanates with computational x-ray absorption spectroscopy

Resolving the local structure motifs that characterize phase evolution as a function of composition is a key challenge in structure characterization of complex materials. Here, in this study, we combine first-principles simulations and x-ray absorption near-edge structures (XANES) analysis to gain insights into the structure evolution revealed by measurements across a combinatorial zinc titanate thin film, which was grown with smoothly varying composition over a wide range of the Ti:Zn ratio. Specifically, we propose a cluster blind-signal-separation (cBSS) method for XANES spectral analysis based on a library of the structures and spectra of representative local motifs. In addition to motifs from zinc titanate crystals, two types of Ti-defect models constructed in this study are key to the understanding of the structure characteristics in the Zn-rich region. The cBSS method makes use of both spectral clustering of the simulated site-XANES spectra library and the BSS procedure to construct high-fidelity spectral basis functions from an experimental spectral sequence. The method provides a rigorous measure of the spectral sensitivity and basis completeness. The results of the XANES analysis are corroborated with other experimental modalities, including x-ray diffraction and spectroscopic ellipsometry, to validate the cBSS method. The calculated motif weights resulting from fitting the XANES spectra with the cBSS basis probe the atomic structure characteristics of both crystalline and amorphous phases as a function of the Ti/Zn composition. The insights of the local structure motif evolution are pivotal to the understanding of the nonmonotonic trend in the optical gap, which may lead to potential applications through tuning the optical properties of zinc titanate. The workflow of the XANES spectral analysis developed in this work can be generalized to construct the structure-property relationship in a broad material space.

36 MATERIALS SCIENCE↗