SEARCH · Search NASA

Results for “data reduction algorithms”

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.

SourceResearch connectionFrom yearThrough year

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

At least 505 records · Page 28

All sources NASA NTRS DOE OSTI

NASA NTRS1989

A novel multistage estimation of the signal parameters of a possibly data-modulated sinusoid under very high dynamics

A novel multistage estimation scheme is presented for estimating the parameters of a received carrier signal possibly phase-modulated by unknown data, and experiencing very high Doppler, Doppler rate, etc. Such a situation arises, for example, in the case of Global Positioning Systems (GPS) where the signal parameters are directly related to the position, velocity, acceleration and jerk of the GPS receiver. In the proposed multistage scheme, the first stage estimator operates as a coarse estimator resulting in higher rms estimation errors but with a relatively small probability of the frequency estimation error exceeding one-half of the sampling frequency (an event termed cycle slip). The second state of the estimator operates on the error signal available from the first stage, refining the overall estimates, and in the process also reduces the number of cycle slips. The first stage algorithm is selected to be a modified least squares algorithm operating upon the differential signal model and referred to as differential least squares. This estimation stage provides relatively coarse estimates of the frequency and its derivatives. The second algorithm is simply an extended Kalman filter which also yields the estimate of the phase along with a more refined estimate of frequency as well. A major advantage of the proposed algorithm is a reduction in the threshold on received carrier power-to-noise power spectral density ratio as compared to the threshold achievable by either of these algorithms alone. In fact, it appears from the simulations that for the case of an unmodulated carrier, the proposed scheme achieves the same threshold as for an almost exact and computationally intensive implementation of the maximum likelihood estimator.

Kumar, Rajendra↗

NASA NTRS2016

Reducing Multisensor Satellite Monthly Mean Aerosol Optical Depth Uncertainty: 1. Objective Assessment of Current AERONET Locations

Various space-based sensors have been designed and corresponding algorithms developed to retrieve aerosol optical depth (AOD), the very basic aerosol optical property, yet considerable disagreement still exists across these different satellite data sets. Surface-based observations aim to provide ground truth for validating satellite data; hence, their deployment locations should preferably contain as much spatial information as possible, i.e., high spatial representativeness. Using a novel Ensemble Kalman Filter (EnKF)- based approach, we objectively evaluate the spatial representativeness of current Aerosol Robotic Network (AERONET) sites. Multisensor monthly mean AOD data sets from Moderate Resolution Imaging Spectroradiometer, Multiangle Imaging Spectroradiometer, Sea-viewing Wide Field-of-view Sensor, Ozone Monitoring Instrument, and Polarization and Anisotropy of Reflectances for Atmospheric Sciences coupled with Observations from a Lidar are combined into a 605-member ensemble, and AERONET data are considered as the observations to be assimilated into this ensemble using the EnKF. The assessment is made by comparing the analysis error variance (that has been constrained by ground-based measurements), with the background error variance (based on satellite data alone). Results show that the total uncertainty is reduced by approximately 27% on average and could reach above 50% over certain places. The uncertainty reduction pattern also has distinct seasonal patterns, corresponding to the spatial distribution of seasonally varying aerosol types, such as dust in the spring for Northern Hemisphere and biomass burning in the fall for Southern Hemisphere. Dust and biomass burning sites have the highest spatial representativeness, rural and oceanic sites can also represent moderate spatial information, whereas the representativeness of urban sites is relatively localized. A spatial score ranging from 1 to 3 is assigned to each AERONET site based on the uncertainty reduction, indicating its representativeness level.

aerosol optical depth↗

NASA NTRS2010

Assessment of Biases in MODIS Surface Reflectance Due to Lambertian Approximation

Using MODIS data and the AERONET-based Surface Reflectance Validation Network (ASRVN), this work studies errors of MODIS atmospheric correction caused by the Lambertian approximation. On one hand, this approximation greatly simplifies the radiative transfer model, reduces the size of the look-up tables, and makes operational algorithm faster. On the other hand, uncompensated atmospheric scattering caused by Lambertian model systematically biases the results. For example, for a typical bowl-shaped bidirectional reflectance distribution function (BRDF), the derived reflectance is underestimated at high solar or view zenith angles, where BRDF is high, and is overestimated at low zenith angles where BRDF is low. The magnitude of biases grows with the amount of scattering in the atmosphere, i.e., at shorter wavelengths and at higher aerosol concentration. The slope of regression of Lambertian surface reflectance vs. ASRVN bidirectional reflectance factor (BRF) is about 0.85 in the red and 0.6 in the green bands. This error propagates into the MODIS BRDF/albedo algorithm, slightly reducing the magnitude of overall reflectance and anisotropy of BRDF. This results in a small negative bias of spectral surface albedo. An assessment for the GSFC (Greenbelt, USA) validation site shows the albedo reduction by 0.004 in the near infrared, 0.005 in the red, and 0.008 in the green MODIS bands.

Wang, Yujie↗

NASA NTRS2005

Evaluation of Airborne Precision Spacing in a Human-in-the-Loop Experiment

A significant bottleneck in the current air traffic system occurs at the runway. Expanding airports and adding new runways will help solve this problem; however, this comes with significant costs: financially, politically and environmentally. A complementary solution is to safely increase the capacity of current runways. This can be achieved by precisely spacing aircraft at the runway threshold, with a resulting reduction in the spacing bu er required under today s operations. At NASA's Langley Research Center, the Airspace Systems program has been investigating airborne technologies and procedures that will assist the flight crew in achieving precise spacing behind another aircraft. A new spacing clearance allows the pilot to follow speed cues from a new on-board guidance system called Airborne Merging and Spacing for Terminal Arrivals (AMSTAR). AMSTAR receives Automatic Dependent Surveillance-Broadcast (ADS-B) reports from an assigned, leading aircraft and calculates the appropriate speed for the ownship to fly to achieve the desired spacing interval, time- or distance-based, at the runway threshold. Since the goal is overall system capacity, the speed guidance algorithm is designed to provide system-wide benefits and stability to a string of arriving aircraft. An experiment was recently performed at the NASA Langley Air Traffic Operations Laboratory (ATOL) to test the flexibility of Airborne Precision Spacing operations under a variety of operational conditions. These included several types of merge and approach geometries along with the complementary merging and in-trail operations. Twelve airline pilots and four controllers participated in this simulation. Performance and questionnaire data were collected from a total of eighty-four individual arrivals. The pilots were able to achieve precise spacing with a mean error of 0.5 seconds and a standard deviation of 4.7 seconds. No statistically significant di erences in spacing performance were found between in-trail and merging operations or among the three modeled airspaces. Questionnaire data showed general acceptance for both pilots and controllers. These results reinforce previous findings from full-mission simulation and flight evaluation of the in-trail operations. This paper reviews the results of this simulation in detail.

Barmore, Bryan E.↗

DOE OSTI2025

TorchBraid: High-Performance Layer-Parallel Training of Deep Neural Networks with MPI and GPU Acceleration

TorchBraid is a high-performance implementation of layer-parallel training for deep neural networks (DNNs) supporting MPI-based parallelism and GPU acceleration. Layer-parallel training has been developed to overcome the serialization inherent in forward and backward propagation of DNNs that limits utilization of computational resources in the strong scaling limit. To achieve this, TorchBraid integrates the PyTorch neural network framework with the state-of-the-art XBraid time-parallel library. Furthermore, this article presents the use and performance of TorchBraid, in addition to solutions for overcoming the algorithmic challenges inherent in combining automatic differentiation with layer-parallel. Results are presented with and without GPU acceleration for the Tiny ImageNet and MNIST image classification data sets, as well as recurrent neural networks. Overall, TorchBraid enables fast training of DNNs, both in a strong and weak scaling context. In addition to the TorchBraid software, several new advances in applying layer-parallel algorithms are detailed. Integration of layer-parallel with data-parallel algorithms is presented for the first time, showing the computational advantages of the combination. Standard deep learning techniques, like batch-normalization, are developed for layer-parallel training. Finally, a new approach combining layer-parallel with spatial coarsening in order to accelerate training for 3D image classification shows roughly a 10× speedup over serial execution.

Layer-parallel↗

NASA NTRSDate not supplied

Enhanced Monitoring and Forecasting of South Asian Air Quality Episodes with Multi-Sensor Satellite Products and Dispersion Modeling

Air pollution and particulates represent a serious environmental, public health, and overarching societal concern in the Hindu Kush Himalaya (HKH) region of south-central Asia, especially during the boreal winter and spring months. Frequent contributors to poor air quality of the region include dust and particulate matter transport from the Middle East region and western India, persistent nocturnal fog and smog during the stable dry monsoon months, and biomass burning during the pre-monsoon months of boreal spring. Nocturnal and multi-day persistent stable fog and smog events can occur during the stable conditions of the dry monsoon months that lead to hazardous visibility reductions (e.g., at major airports such as Delhi, India)and poor air quality, particularly in urban corridors.Our project goal is therefore to design and implement a robust air quality and chemistry observation and modeling system using the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) and the Weather Research and Forecasting coupled with Chemistry (WRF-Chem) models. This system effectively assimilates aerosol and trace gas retrievals from geostationary and polar-orbiting satellites to advance the monitoring and prediction capabilities of air quality and visibility reductions throughout the HKH region. For this presentation, we highlight several[multi-spectral] geostationary satellite products and dispersion modeling capabilities being developed and implemented for improved situational awareness and forecasting of air quality episodes.Launched in December 2018, the Advanced Meteorological Imager (AMI) aboard the Geostationary-Korean Multi-Purpose Satellite-2A (GEO-KOMPSAT-2A, or GK2A) provides 16 channels of multi-spectral information similar to the U.S. GOES-16/17 satellites. Proven community recipes for generating multi-spectral Red-Green-Blue (RGB) composite products to detect specific meteorological phenomena are applied to the AMI data for the HKH region. We implemented four such RGB products consisting of truecolor with Rayleigh correction to aid in smoke detection; dust RGB for monitoring larger particulate matter; nighttime microphysics for fog/smog detection;and natural color fire RGB and the shortwave 3.8 micron channel for fire hot-spot detection.We additionally configured HYSPLIT for dust/sand dispersion and concentration forecasts over HKH. The HYSPLIT dispersion simulations are initially being conducted with two different methodologies: (1) dust/sand lofting based on initializing plume releases from visual inspection of the dust RGB imagery; and (2) dust/sand release and dispersion using the internal HYSPLIT algorithm that determines lofting from input wind speeds and prescribed land use. The goal is to develop a near real-time HYSPLIT modeling solution for routinely simulating the transport and dispersion of dust plumes across the HKH region. This presentation illustrates these products and capabilities for various use cases from the 2019 to 2021 period, highlighting a particularly unhealthy episode during late March 2021 across Nepal.

Remote Sensing↗

NASA NTRS2005

NASA Tech Briefs, December 2005

Topics covered include: Video Mosaicking for Inspection of Gas Pipelines; Shuttle-Data-Tape XML Translator; Highly Reliable, High-Speed, Unidirectional Serial Data Links; Data-Analysis System for Entry, Descent, and Landing; Hybrid UV Imager Containing Face-Up AlGaN/GaN Photodiodes; Multiple Embedded Processors for Fault-Tolerant Computing; Hybrid Power Management; Magnetometer Based on Optoelectronic Microwave Oscillator; Program Predicts Time Courses of Human/ Computer Interactions; Chimera Grid Tools; Astronomer's Proposal Tool; Conservative Patch Algorithm and Mesh Sequencing for PAB3D; Fitting Nonlinear Curves by Use of Optimization Techniques; Tool for Viewing Faults Under Terrain; Automated Synthesis of Long Communication Delays for Testing; Solving Nonlinear Euler Equations With Arbitrary Accuracy; Self-Organizing-Map Program for Analyzing Multivariate Data; Tool for Sizing Analysis of the Advanced Life Support System; Control Software for a High-Performance Telerobot; Java Radar Analysis Tool; Architecture for Verifiable Software; Tool for Ranking Research Options; Enhanced, Partially Redundant Emergency Notification System; Close-Call Action Log Form; Task Description Language; Improved Small-Particle Powders for Plasma Spraying; Bonding-Compatible Corrosion Inhibitor for Rinsing Metals; Wipes, Coatings, and Patches for Detecting Hydrazines; Rotating Vessels for Growing Protein Crystals; Oscillating-Linear-Drive Vacuum Compressor for CO2; Mechanically Biased, Hinged Pairs of Piezoelectric Benders; Apparatus for Precise Indium-Bump Bonding of Microchips; Radiation Dosimetry via Automated Fluorescence Microscopy; Multistage Magnetic Separator of Cells and Proteins; Elastic-Tether Suits for Artificial Gravity and Exercise; Multichannel Brain-Signal-Amplifying and Digitizing System; Ester-Based Electrolytes for Low-Temperature Li-Ion Cells; Hygrometer for Detecting Water in Partially Enclosed Volumes; Radio-Frequency Plasma Cleaning of a Penning Malmberg Trap; Reduction of Flap Side Edge Noise - the Blowing Flap; and Preventing Accidental Ignition of Upper-Stage Rocket Motors.

Source record↗

NASA NTRS2017

Colloid Microthruster Flight Performance Results from Space Technology 7 Disturbance Reduction System

Space Technology 7 Disturbance Reduction System (ST7-DRS) is a NASA technology demonstration payload as part of the ESA LISA Pathfinder (LPF) mission, which launched on December 3, 2015. The ST7-DRS payload includes colloid microthrusters as part of a drag-free dynamic control system (DCS) hosted on an integrated avionics unit (IAU) with spacecraft attitude and test mass position provided by the LPF spacecraft computer and the highly sensitive gravitational reference sensor (GRS) as part of the LISA Technology Package (LTP). The objective of the DRS was to validate two technologies: colloid micro-Newton thrusters (CMNT) to provide low-noise control capability of the spacecraft, and drag-free flight control. The CMNT were developed by Busek Co., Inc., in a partnership with NASA Jet Propulsion Laboratory (JPL), and the DCS algorithms and flight software were developed at NASA Goddard Space Flight Center (GSFC). ST7-DRS demonstrated drag-free operation with 10nmHz level precision spacecraft position control along the primary axis of the LTP using eight CMNTs that provided 5-30 N each with 0.1 N precision. The DCS and CMNTs performed as required and as expected from ground test results, meeting all Level 1 requirements based on on-orbit data and analysis. DRS microthrusters operated for 2400 hours in flight during commissioning activities, a 90-day experiment and the extended mission. This mission represents the first validated demonstration of electrospray thrusters in space, providing precision spacecraft control and drag-free operation in a flight environment with applications to future gravitational wave observatories like LISA.

Ziemer, John↗

NASA NTRS2002

Spectral Re-Growth Reduction for CCSDS 8-D 8-PSK TCM

This report presents a study on the CCSDS recommended 8-dimensional 8 PSK Trellis Coded Modulation (TCM) scheme. The important steps of the CCSDS scheme include: conversion of serial data into parallel form, differential encoding, convolutional encoding, constellation mapping, and filtering the 8-PSK symbols using the square root raised cosine (SRRC) pulses. The last step, namely the filtering of the 8 PSK symbols using SRRC pulses, significantly affects the bandwidth of the signal. If a nonlinear power amplifier is used, the SRRC filtered signal creates spectral regrowth. The purpose of this report is to investigate a technique, called the smooth phase interpolated keying (SPIK), that can provide an alternative to SRRC filtering so that good spectral as well as power efficiencies can be obtained with the CCSDS encoder. The results of this study show that the CCSDS encoder does not affect the spectral shape of the SRRC filtered signal or the SPIK signal. When a nonlinear traveling wave tube amplifier (TWTA) is used, the spectral performance of the SRRC signal degrades significantly while the spectral performance of SPIK remains unaffected. The degrading effect of a nonlinear solid state power amplifier (SSPA) on SRRC is found to be less than that due to a nonlinear TWTA. However, in both cases, the spectral performance of the SRRC modulated signal is worse than that of the SPIK signal. The bit error rate (BER) performance of the SRRC signal in a linear amplifier environment is about 2.5 dB better than that of the SPIK signal when both the receivers use algorithms of similar complexity. In a nonlinear TWTA environment, the SRRC signal requires accurate phase tracking since the TWTA introduces additional phase distortion. This problem does not arise with SPIK signal due to its constant envelope property. When a nonlinear amplifier is used, the SRRC method loses nearly 1 dB in the bit error rate performance. The SPIK signal does not lose any performance. Thus the performance gap between SRRC and SPIK reduces. The BER performance of SPIK can be improved even further by using a more optimal receiver. A similar optimal receiver for SRRC is quite complex since the amplifier distorts the pulse shape. However, this requires further investigation and is not covered in this report.

Borah, Deva K.↗

NASA NTRS2024

Parameterizing Spectral Surface Reflectance Relationships for the Dark Target Aerosol Algorithm Applied to A Geostationary Imager

Originally developed for the Moderate Resolution Imaging Spectroradiometer (MODIS) in polar, sun-synchronous low-earth orbit (LEO), the Dark Target (DT) aerosol retrieval algorithm relies on the assumption of a Surface Reflectance Parameterization (SRP) over land surfaces. Specifically for vegetated and dark-soiled surfaces, values of surface reflectance in blue and red visible-wavelength bands are assumed to be nearly linearly related to each other and to the value in a shortwave infrared (SWIR) wavelength band. This SRP also includes dependencies on scattering angle and a normalized difference vegetation index computed from two SWIR bands (NDVISWIR). As the DT retrieval algorithm is being ported to new sensors to continue and expand the aerosol data record, we assess whether the MODIS-assumed SRP can be used for these sensors. Here, we specifically assess SRP for the Advanced Baseline Imager (ABI) aboard, the Geostationary Operational Environmental Satellite (GOES)-16/East (ABIE). First, we find that using MODIS-based SRP leads to higher biases and artificial diurnal signatures in aerosol optical depth (AOD)retrievals from ABIE. The primary reason appears to be that geostationary orbit (GEO) encounters an entirely different set of observation geometry than does LEO, primarily with regards to solar angles coupled with fixed view angles. Therefore, we have developed a new SRP for GEO that draws the angular shape of the surface bidirectional reflectance. We also introduce modifications to the parametrization of both red-SWIR and blue-red spectral relationships to include additional information.The revised Red-SWIR SRP includes solar zenith angle, NDVISWIR, and land-type percentage from an ancillary database. The blue-red SRP adds dependencies on the scattering angle and NDVISWIR. The new SRPs improve the AOD retrieval of ABIE in terms of overall less bias and mitigation of the overestimation around local noon. The average bias of DT AOD compared to AERONET AOD shows a reduction from 0.08 to 0.03, while the bias of local solar noon decreases from 0.12 to 0.03.The agreement between DT and AERONET AOD is established through regression slope of 1.06 and y-intercept of 0.01with correlation coefficient is0.74. By using the new SRP, the percentage of data falling within the expected error range (±0.05 + 15%) is notably risen from 54% to 78%.

Mijin Kim↗

NASA NTRS2020

Amino Acid Encoding for Deep Learning Applications

Background: The number of applications of deep learning algorithms in bioinformatics is increasing as they usually achieve superior performance over classical approaches, especially, when bigger training datasets are available. In deep learning applications, discrete data, e.g. words or n-grams in language, or amino acids or nucleotides in bioinformatics, are generally represented as a continuous vector through an embedding matrix. Recently, learning this embedding matrix directly from the data as part of the continuous iteration of the model to optimize the target prediction – a process called ‘end-to-end learning’ – has led to state-ofthe-art results in many fields. Although usage of embeddings is well described in the bioinformatics literature, the potential of end-to-end learning for single amino acids, as compared to more classical manually-curated encoding strategies, has not been systematically addressed. To this end, we compared classical encoding matrices, namely one-hot, VHSE8 and BLOSUM62, to end-to-end learning of amino acid embeddings for two different prediction tasks using three widely used architectures, namely recurrent neural networks (RNN), convolutional neural networks (CNN), and the hybrid CNN-RNN. Results: By using different deep learning architectures, we show that end-to-end learning is on par with classical encodings for embeddings of the same dimension even when limited training data is available, and might allow for a reduction in the embedding dimension without performance loss, which is critical when deploying the models to devices with limited computational capacities. We found that the embedding dimension is a major factor in controlling the model performance. Surprisingly, we observed that deep learning models are capable of learning from random vectors of appropriate dimension. Conclusion: Our study shows that end-to-end learning is a flexible and powerful method for amino acid encoding. Further, due to the flexibility of deep learning systems, amino acid encoding schemes should be benchmarked against random vectors of the same dimension to disentangle the information content provided by the encoding scheme from the distinguishability effect provided by the scheme.

Hesham ElAbd↗

DOE OSTI2025

Integration of a grey-box refrigerated case model in EnergyPlus via Python plugin

Commercial buildings, in particular grocery stores (due mainly to their large refrigeration load), provide opportunities for energy cost reductions. Grocery stores could offer substantial load flexibility to the power grid through participation in demand response programs because of their usage patterns and relatively high energy intensity. This load flexibility could come from modifying the control of heating, ventilation, and air conditioning (HVAC) systems, refrigeration systems, or both. Although estimation of the HVAC system’s load flexibility potential is relatively targeted in the literature, estimating load flexibility of refrigeration systems is nascent and has been a challenge, in part because of the lack of proper simulation tools that capture the dynamics in the refrigeration cases. The existing refrigerated case model within EnergyPlus, a whole building energy simulation program, assumes a constant case temperature throughout the simulation period and does not explicitly model the cycling of the compressor serving the refrigerated case. In addition, it does not encompass modeling of temperatures of the product inside the refrigerated case. This difference between modeled and actual operation can be a barrier to the development of demand control algorithm and accurate analysis of load flexibility potential. In this paper, we present a grey-box model for modeling refrigerated cases in grocery stores, which include medium temperature and low temperature. Four cases are modeled; two are low-temperature closed cases and two are medium-temperature cases with one closed and one open. Data from an experimental facility are used to train and test the models. Results demonstrate the efficacy of the grey-box models in predicting the temperatures. This model is integrated into EnergyPlus to capture the dynamic effects of case temperature on the environment and enhance the calculation of sensible and latent heat exchange with the environment (case credits). These enhancements can be leveraged more broadly to model advanced refrigeration controls such as defrost, develop and test unique algorithms that could affect refrigeration interactions with HVAC, and refine store design for any commercial building with refrigeration.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

DOE OSTI2026

UCB-GLOBES: An open-access mass spectral database of identified and unidentified atmospheric organic compounds

Chemical characterization of atmospheric organic aerosols using gas chromatography with 70 eV electron ionization mass spectrometry (GC/EI-MS) has been used for decades in advancing molecular marker detection and identification, though primarily through suspect screening and/or targeted analyses. To advance non-targeted analyses of environmental samples, we have catalogued approximately 27 000 mass spectra (MS) of the trimethylsilyl derivatives of semi-volatile organic aerosol (OA) analytes in the open-access University of California Berkeley Goldstein Library of Organic Biogenic Environmental Spectra (UCB-GLOBES). Analytes were observed in ambient samples from the U.S. and the Central Amazon and/or laboratory simulations of secondary OA (SOA) formation. These samples are representative of OA under urban and biomass burning influences as well as SOA derived from biogenic precursors (e.g., isoprene, monoterpenes, sesquiterpenes) and biomass burning intermediates. MS are documented in UCB-GLOBES without regard to known chemical identity, annotated with extensive metadata such as sample source/experimental conditions, any structural information gained from MS analyses, and predicted chemical properties such as average carbon oxidation state and carbon number. UCB-GLOBES MS are compatible for importing into the NIST MS Search program, and we have also provided a Jupyter Notebook for MS visualization and comparisons. We demonstrate the utility of UCB-GLOBES through MS reanalyses of prior analytes observed in ambient data, finding a 20 % reduction in the number of analytes assigned to OA source categories reliant solely on time series correlation and an overall 11 % increase in new MS-based OA source categorization for the Southeast U.S. For 1513 analytes observed previously in the Central Amazon, we found 375 MS matches using UCB-GLOBES vs. 136 MS matches during prior analyses, representing a 14 % gain in newly confirmed or newly categorized OA species. While OA from laboratory oxidation experiments in UCB-GLOBES are highly diverse chemically, on average only 29 % of UCB-GLOBES MS have a mass spectral match to another MS entry in UCB-GLOBES and/or in databases of known compounds (i.e. NIST MS Database, Adams Essential Oil, MANE Flavor and Fragrance Company). This indicates that roughly 70 % of UCB-GLOBES MS are unique thus far, not observed more than once among the laboratory oxidation samples and ambient data in UCB-GLOBES MS. Further, only 18 % can be positively identified using these databases or known authentic standards. This points to a large gap between these laboratory simulations and ambient OA. Overall, the UCB-GLOBES database can be utilized for improving confidence in OA source categorization and/or identification, novel chemical marker discovery, tracking chemical diversity, de novo structure and properties prediction, and improving MS search and matching algorithms. This can ultimately inform future research priorities for the chemical characterization of atmospheric organic samples.

Mass spectrometry↗

NASA NTRS2020

Amino Acid Encoding for Deep Learning Applications

Background: The number of applications of deep learning algorithms in bioinformatics is increasing as they usually achieve superior performance over classical approaches, especially, when bigger training datasets are available. In deep learning applications, discrete data, e.g. words or n-grams in language, or amino acids or nucleotides in bioinformatics, are generally represented as a continuous vector through an embedding matrix. Recently, learning this embedding matrix directly from the data as part of the continuous iteration of the model to optimize the target prediction – a process called ‘end-to-end learning’ – has led to state-of-the-art results in many fields. Although usage of embeddings is well described in the bioinformatics literature, the potential of end-to-end learning for single amino acids, as compared to more classical manually-curated encoding strategies, has not been systematically addressed. To this end, we compared classical encoding matrices, namely one-hot, VHSE8 and BLOSUM62, to end-to-end learning of amino acid embeddings for two different prediction tasks using three widely used architectures, namely recurrent neural networks (RNN), convolutional neural networks (CNN), and the hybrid CNN-RNN. Results: By using different deep learning architectures, we show that end-to-end learning is on par with classical encodings for embeddings of the same dimension even when limited training data is available, and might allow for a reduction in the embedding dimension without performance loss, which is critical when deploying the models to devices with limited computational capacities. We found that the embedding dimension is a major factor in controlling the model performance. Surprisingly, we observed that deep learning models are capable of learning from random vectors of appropriate dimension. Conclusion: Our study shows that end-to-end learning is a flexible and powerful method for amino acid encoding. Further, due to the flexibility of deep learning systems, amino acid encoding schemes should be benchmarked against random vectors of the same dimension to disentangle the information content provided by the encoding scheme from the distinguishability effect provided by the scheme.

Deep-learning↗

NASA NTRS2013

RUSHMAPS: Real-Time Uploadable Spherical Harmonic Moment Analysis for Particle Spectrometers

RUSHMAPS is a new onboard data reduction scheme that gives real-time access to key science parameters (e.g. moments) of a class of heliophysics science and/or solar system exploration investigation that includes plasma particle spectrometers (PPS), but requires moments reporting (density, bulk-velocity, temperature, pressure, etc.) of higher-level quality, and tolerates a lowpass (variable quality) spectral representation of the corresponding particle velocity distributions, such that telemetry use is minimized. The proposed methodology trades access to the full-resolution velocity distribution data, saving on telemetry, for real-time access to both the moments and an adjustable-quality (increasing quality increases volume) spectral representation of distribution functions. Traditional onboard data storage and downlink bandwidth constraints severely limit PPS system functionality and drive cost, which, as a consequence, drives a limited data collection and lower angular energy and time resolution. This prototypical system exploit, using high-performance processing technology at GSFC (Goddard Space Flight Center), uses a SpaceCube and/or Maestro-type platform for processing. These processing platforms are currently being used on the International Space Station as a technology demonstration, and work is currently ongoing in a new onboard computation system for the Earth Science missions, but they have never been implemented in heliospheric science or solar system exploration missions. Preliminary analysis confirms that the targeted processor platforms possess the processing resources required for realtime application of these algorithms to the spectrometer data. SpaceCube platforms demonstrate that the target architecture possesses the sort of compact, low-mass/power, radiation-tolerant characteristics needed for flight. These high-performing hybrid systems embed unprecedented amounts of onboard processing power in the CPU (central processing unit), FPGAs (field programmable gate arrays), and DSP (digital signal processing) elements. The fundamental computational algorithm de constructs 3D velocity distributions in terms of spherical harmonic spectral coefficients (which are analogous to a Fourier sine-cosine decomposition), but uses instead spherical harmonics Legendre polynomial orthogonal functions as a basis for the expansion, portraying each 2D angular distribution at every energy or, geometrically, spherical speed-shell swept by the particle spectrometer. Optionally, these spherical harmonic spectral coefficients may be telemetered to the ground. These will provide a smoothed description of the velocity distribution function whose quality will depend on the number of coefficients determined. Successfully implemented on the GSFC-developed processor, the capability to integrate the proposed methodology with both heritage and anticipated future plasma particle spectrometer designs is demonstrated (with sufficiently detailed design analysis to advance TRL) to show specific science relevancy with future HSD (Heliophysics Science Division) solar-interplanetary, planetary missions, sounding rockets and/or CubeSat missions.

Figueroa-Vinas, Adolfo↗

NASA NTRS2012

NASA's Robotic Lunar Lander Development Project

Since early 2005, NASA's Robotic Lunar Lander Development (RLLD) office at NASA MSFC, in partnership with the Applied Physics Laboratory (APL), has developed mission concepts and preformed risk-reduction activities to address planetary science and exploration objectives uniquely met with landed missions. The RLLD team developed several concepts for lunar human-exploration precursor missions to demonstrate precision landing and in-situ resource utilization, a multi-node lunar geophysical network mission, either as a stand-alone mission, or as part of the International Lunar Network (ILN), a Lunar Polar Volatiles Explorer and a Mercury lander mission for the Planetary Science decadal survey, and an asteroid rendezvous and landing mission for the Exploration Precursor Robotics Mission (xPRM) office. The RLLD team has conducted an extensive number of risk-reduction activities in areas common to all lander concepts, including thruster testing, propulsion thermal control demonstration, composite deck design and fabrication, and landing leg stability and vibration. In parallel, the team has developed two robotic lander testbeds providing closed-loop, autonomous hover and descent activities for integration and testing of flight-like components and algorithms. A compressed-air test article had its first flight in September 2009 and completed over 150 successful flights. This small test article (107 kg dry/146 kg wet) uses a central throttleable thruster to offset gravity, plus 3 descent thrusters (~37lbf ea) and 6 attitude-control thrusters (~12lbf ea) to emulate the flight system with pulsed operation over approximately 10s of flight time. The test article uses carbon composite honeycomb decks, custom avionics (COTS components assembled in-house), and custom flight and ground software. A larger (206 kg dry/322 kg wet), hydrogen peroxide-propelled vehicle began flight tests in spring 2011 and fly over 30 successful flights to a maximum altitude of 30m. The monoprop testbed also uses a central gravity-canceling thruster and 3 descent thrusters, but has 12 attitude-control thrusters and a maximum flight time of over a minute. The testbed uses aluminum ortho-grid decks, an LN200-1 IMU, Roke Manor Radar Altimeter, Illunis optical cameras, Novatel Pro-Pak GPS truth data system, Pressure transducers & thermocouples for housekeeping, "In-Control" ground system software, and the core Flight Executive (cFE) modular software environment. The peroxide lander testbed is able to accept other sensors and algorithms for testing, both from within NASA and from other customers. Through these activities, the RLLD team has significantly reduced technical risks for all small and medium class robotic landers for the Moon and other airless planetary bodies.

Cohen, Barbara A.↗

NASA NTRS2016

Model-Based Fatigue Prognosis of Fiber-Reinforced Laminates Exhibiting Concurrent Damage Mechanisms

Prognostics of large composite structures is a topic of increasing interest in the field of structural health monitoring for aerospace, civil, and mechanical systems. Along with recent advancements in real-time structural health data acquisition and processing for damage detection and characterization, model-based stochastic methods for life prediction are showing promising results in the literature. Among various model-based approaches, particle-filtering algorithms are particularly capable in coping with uncertainties associated with the process. These include uncertainties about information on the damage extent and the inherent uncertainties of the damage propagation process. Some efforts have shown successful applications of particle filtering-based frameworks for predicting the matrix crack evolution and structural stiffness degradation caused by repetitive fatigue loads. Effects of other damage modes such as delamination, however, are not incorporated in these works. It is well established that delamination and matrix cracks not only co-exist in most laminate structures during the fatigue degradation process but also affect each other's progression. Furthermore, delamination significantly alters the stress-state in the laminates and accelerates the material degradation leading to catastrophic failure. Therefore, the work presented herein proposes a particle filtering-based framework for predicting a structure's remaining useful life with consideration of multiple co-existing damage-mechanisms. The framework uses an energy-based model from the composite modeling literature. The multiple damage-mode model has been shown to suitably estimate the energy release rate of cross-ply laminates as affected by matrix cracks and delamination modes. The model is also able to estimate the reduction in stiffness of the damaged laminate. This information is then used in the algorithms for life prediction capabilities. First, a brief summary of the energy-based damage model is provided. Then, the paper describes how the model is embedded within the prognostic framework and how the prognostics performance is assessed using observations from run-to-failure experiments

Prognostics↗

NASA NTRS1994

An efficient system for reliably transmitting image and video data over low bit rate noisy channels

This research project is intended to develop an efficient system for reliably transmitting image and video data over low bit rate noisy channels. The basic ideas behind the proposed approach are the following: employ statistical-based image modeling to facilitate pre- and post-processing and error detection, use spare redundancy that the source compression did not remove to add robustness, and implement coded modulation to improve bandwidth efficiency and noise rejection. Over the last six months, progress has been made on various aspects of the project. Through our studies of the integrated system, a list-based iterative Trellis decoder has been developed. The decoder accepts feedback from a post-processor which can detect channel errors in the reconstructed image. The error detection is based on the Huber Markov random field image model for the compressed image. The compression scheme used here is that of JPEG (Joint Photographic Experts Group). Experiments were performed and the results are quite encouraging. The principal ideas here are extendable to other compression techniques. In addition, research was also performed on unequal error protection channel coding, subband vector quantization as a means of source coding, and post processing for reducing coding artifacts. Our studies on unequal error protection (UEP) coding for image transmission focused on examining the properties of the UEP capabilities of convolutional codes. The investigation of subband vector quantization employed a wavelet transform with special emphasis on exploiting interband redundancy. The outcome of this investigation included the development of three algorithms for subband vector quantization. The reduction of transform coding artifacts was studied with the aid of a non-Gaussian Markov random field model. This results in improved image decompression. These studies are summarized and the technical papers included in the appendices.

Costello, Daniel J., Jr.↗