Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Training Time”

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.

At least 397 records · Page 22
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Fully Convolutional Spatio-Temporal Models for Representation Learning in Plasma Science

We have trained a fully convolutional spatio-temporal model for fast and accurate representation learning in the challenging exemplar application area of fusion energy plasma science. The onset of major disruptions is a critically important fusion energy science issue that must be resolved for advanced tokamak plasmas such as the $25B burning plasma international thermonuclear experimental reactor (ITER) experiment. While a variety of statistical methods have been used to address the problem of tokamak disruption prediction and control, recent approaches based on deep learning have proven particularly compelling. In the present paper, we introduce further improvements to the fusion recurrent neural network (FRNN) software suite, which delivered cross-machine disruption predictions with unprecedented accuracy using a large database of experimental signals from two major tokamaks. Up to now, FRNN was based on the long short-term memory (LSTM) variant of recurrent neural networks to leverage the temporal information in the data. Here, we implement and apply the "temporal convolutional neural network (TCN)" architecture to the time-dependent input signals. Furthermore, this allows highly optimized convolution operations to carry the majority of the computational load of training, thus enabling a reduction in training time, and the effective use of high-performance computing resources for hyperparameter tuning. At the same time, the TCN-based architecture achieves better predictive performance when compared with the LSTM architecture for various tasks for a representative fusion database.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2020

Machine learning a molecular Hamiltonian for predicting electron dynamics

We develop a computational method to learn a molecular Hamiltonian matrix from matrix-valued time series of the electron density. As we demonstrate for three small molecules, the resulting Hamiltonians can be used for electron density evolution, producing highly accurate results even when propagating 1,000 time steps beyond the training data. As a more rigorous test, we use the learned Hamiltonians to simulate electron dynamics in the presence of an applied electric field, extrapolating to a problem that is beyond the field-free training data. We find that the resulting electron dynamics predicted by our learned Hamiltonian are in close quantitative agreement with the ground truth. Our method relies on combining a reduced-dimensional, linear statistical model of the Hamiltonian with a time-discretization of the quantum Liouville equation within time-dependent Hartree Fock theory. In conclusion, we train the model using a least-squares solver, avoiding numerous, CPU-intensive optimization steps. For both field-free and field-on problems, we quantify training and propagation errors, highlighting areas for future development.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
NASA NTRSDate not supplied

Increased Treadmill Running and Lower-Body Resistance Training Intensity May Attenuate Aerobic Capacity Loss During Spaceflight

BACKGROUND: In-flight exercise training is the primary countermeasure to mitigate spaceflight-induced decrements in strength and aerobic capacity. Despite high-quality exercise capabilities and prescriptions aboard the International Space Station (ISS), most astronauts experience strength and aerobic capacity loss; however, significant individual variability exists. Recent findings suggest that greater in-flight running (i.e., distance per session) and lower-body resistance exercise intensity (i.e., load relative to body weight) help preserve strength. Whether similar variables contribute to the preservation of aerobic capacity during spaceflight is unknown. METHODS: A NASA database of exercise performance metrics was used to identify long-duration (≥ 90 days) ISS crew who completed pre- and post-flight aerobic capacity testing (VO2peak, cycle ergometry, n = 44). “Responders” (R) were classified as those that maintained or improved VO2peak (n = 13; 1.4 ± 3.3%), and “non-responders” (NR) were those with the most severe decrements (n = 13; -20.0 ± 3.4%) from pre- to post-flight. In-flight exercise records (cycling, treadmill, resistance exercise) were aggregated across flight weeks. Independent t-tests or Mann-Whitney U tests were used to compare training variables (i.e., frequency, time, intensity, volume) between R and NR. RESULTS: Consistent with the variables contributing to strength protection, in-flight running distance (R: 5.0 ± 0.7; NR: 4.2 ± 1.0 km/session, p = 0.020) and lower-body resistance exercise intensity (R: 110.9 ± 27.3; NR: 91.7 ± 16.5% body weight, p = 0.040) were greater in R than NR. Active running time was also greater in R than NR (R: 101.1 ± 20.4; NR: 82.3 ± 23.5 min/wk; p = 0.039). However, no group differences in cycling frequency, intensity, or time, as well as upper body or core resistance exercise intensity, volume, or frequency were observed. CONCLUSIONS: Although physiologically distinct stimuli, the exercise training variables contributing to the preservation of both strength and aerobic capacity during microgravity exposure appear to be consistent. Increased running time and distance, but not cycling, combined with higher-intensity lower-body resistance exercise, may help mitigate losses in both aerobic capacity and strength during long-duration spaceflight.

Alyssa N Varanoske↗
NASA NTRS2018

Time-Varying Manual Control Identification in a Stall Recovery Task under Different Simulator Motion Conditions

This paper adds data to help develop simulator motion guidelines for stall recovery training by identifying time-varying manual control behavior in a stall recovery task under different simulator motion conditions. A study was conducted in the NASA Ames Vertical Motion Simulator, where seventeen general aviation pilots performed a stall recovery task. Pilots had to follow a flight director through four stages of the stall recovery task. A time-varying identification method was used to quantify how pilots weigh position and velocity information throughout different stages of the task, in both roll and pitch. Four motion configurations were used: no motion, generic hexapod motion, enhanced hexapod motion and full motion. Pilot performance was highest for the enhanced hexapod and full motion conditions in both roll and pitch, and the lowest for the condition with no motion. The time-varying identification method revealed that, in the roll axis, pilot position gain did not significantly change between time segments, but was the lowest for the condition with no motion. The pilot velocity gain was significantly different between motion conditions, the largest difference being found at the beginning of the stall. The enhanced hexapod motion condition had the highest pilot velocity gain. In the pitch axis, the pilot position gain was significantly different between time segments but not between motion conditions. The pitch pilot velocity gain was highest for the full motion condition and increased at the beginning of the stall, but did not change significantly for the other motion conditions. Overall, pilot control behavior under enhanced hexapod motion was more similar to that under full aircraft motion compared to standard hexapod motion. This indicates that motion cueing on hexapod simulators might be improved for stall recovery training by using the enhanced hexapod motion developed in previous experiments.

stall recovery↗
NASA NTRSDate not supplied

Exploring the Moon: Training Astronauts for Artemis Missions

For the first time in over 50 years, NASA scientists are training astronauts assigned to lunar missions. The successful Artemis 1 test flight in 2022 set the stage for future human missions to the Moon. The Artemis II mission will send astronauts to fly by the Moon, providing the first opportunity since Apollo for humans to directly observe and document the lunar surface from a few thousand kilometers, including the lunar far side. Artemis III will be the first human mission to land and explore the Moon’s surface near the lunar South Pole. To prepare for Artemis exploration, geology training for the Artemis II crew and detailed training plans for Artemis III and beyond are in full swing. NASA scientists have been working for several years with the Flight Operations Directorate (FOD) at the Johnson Space Center on a progressive geology training program that starts with incoming astronaut classes, offers opportunities to maintain the geology proficiency between flight assignments, and trains the broader engineering-operations-management community in geology and field operations. In the spring of 2023, our team developed an introductory week-long class called Lunar Fundamentals and delivered this content to the Artemis II crew, as well as key NASA Flight Control Team members. The Lunar Fundamentals curriculum combines basic geographic and geologic information about the Moon; includes detailed content on major geologic processes, the lunar environment, lunar volatiles, and sample science; provides historical context of our growing knowledge about the Moon; and introduces new lunar data sets and tools for exploring the lunar surface. With the Lunar Fundamentals foundation, we continue development of the advanced planetary science classroom content for the Artemis III crew, together with field training science objectives and plans. Partnering with our FOD colleagues, we are building an integrated training plan for lunar science, exploration field methods, and operational skills to ensure that the Artemis astronauts exploring the Moon will have the necessary foundation in geology and lunar science to achieve mission success.

Astronauts training↗
DOE OSTI2024

Robustness of topological persistence in knowledge distillation for wearable sensor data

Topological data analysis (TDA) has shown great success in various applications involving wearable sensor data. However, there are difficulties in leveraging topological features in machine learning and wearable sensors because of the large time consumption and computational resources required to extract the features. To address this problem, knowledge distillation (KD) is utilized to generate a small model and accommodate topological features with persistence image (PI) representations from the raw time series data. Deploying topological knowledge in KD enables the student to achieve better performance compared to the one trained solely on raw time series data. However, it is not yet known if there are coherent characteristics for topological features in PI, which can aid in improving the performance during KD. In this paper, we investigate the suitability and challenges of utilizing topological features in KD for wearable sensor data, thereby contributing to the advancement of the field. Our study explores the impact of transferred topological features by comparing the Teacher-to-Student framework with Multiple Teachers-to-Student where teachers utilize both time series data and persistence images obtained by TDA as inputs. Additionally, we conduct a rigorous examination of topological knowledge effects by testing under various corruptions, knowledge types, and learning strategies in the context of human activity recognition tasks. Our analysis of topological features in KD presents the optimal strategy for incorporating these features. This study includes datasets of varying scales, window lengths, and activity classes, providing a comprehensive evaluation. Our results demonstrate that leveraging topological features in KD to enhance performance across databases.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Evaluating Scalograms for Seismic Event Denoising

Denoising contaminated seismic signals for later processing is a fundamental problem in seismic signals analysis. The most straightforward denoising approach, using spectral filtering, is not effective when noise and seismic signal occupy the same frequency range. Neural network approaches have shown success denoising local signal when trained on short-time Fourier transform spectrograms (Zhu et al 2018; Tibi et al 2021). Scalograms, a wavelet-based transform, achieved ~15% better reconstruction as measured by dynamic time warping on a seismic waveform test set than spectrograms, suggesting their use as an alternative for denoising. We train a deep neural network on a scalogram dataset derived from waveforms recorded by the University of Utah Seismograph Stations network. We find that initial results are no better than a spectrogram approach, with additional overhead imposed by the significantly larger size of scalograms. A robust exploration of neural network hyperparameters and network architecture was not performed, which could be done in follow on work.

58 GEOSCIENCES↗
NASA NTRS1993

Attitude control of spacecraft using neural networks

This paper investigates the use of radial basis function neural networks for adaptive attitude control and momentum management of spacecraft. In the first part of the paper, neural networks are trained to learn from a family of open-loop optimal controls parameterized by the initial states and times-to-go. The trained is then used for closed-loop control. In the second part of the paper, neural networks are used for direct adaptive control in the presence of unmodeled effects and parameter uncertainty. The control and learning laws are derived using the method of Lyapunov.

Vadali, Srinivas R.↗
DOE OSTI2020

Pattern generator circuit for high-speed pulse generation

A pattern generator circuit includes a high-speed shift register and a non-overlap generator. The shift register is programmable to produce a pulse train of pulses having a defined pulse duration and a defined pulse-to-pulse interval. The non-overlap generator deserializes the incoming pulse train, and it produces a time-separated reset pulse based on the pulse train. The shift register is configured to permit the pulse durations and pulse-to-pulse interval to be selected arbitrarily over specified ranges in increments of a basic time unit that depends on the oscillator period.

42 ENGINEERING↗
NASA NTRS2018

Time-Varying Manual Control Identification in a Stall Recovery Task Under Different Simulator Motion Conditions

This paper adds data to help the development of simulator motion cueing guidelines for stall recovery training by identifying time-varying manual control behavior in a stall recovery task under different simulator motion conditions. A study was conducted with seventeen general aviation pilots in the NASA Ames Vertical Motion Simulator. Pilots had to follow a flight director through four stages of a high-altitude stall task. A time-varying identification method was used to quantify how pilot manual control parameters change throughout different stages of the task in both roll and pitch. Four motion configurations were used: no motion, generic hexapod motion, enhanced hexapod motion and full motion. Pilot performance was highest for the enhanced hexapod and full motion configurations in both roll and pitch, and the lowest without motion. In the roll axis, the pilot position gain did not significantly change throughout the stall task, but was the lowest for the condition with no motion. The pilot roll velocity gain was significantly different between motion conditions, the largest difference being found close to the stall point. The enhanced hexapod motion condition had the highest pilot roll velocity gain. In the pitch axis, the pilot position gain was significantly different between time segments but not between motion conditions. The pilot pitch velocity gain was highest for the full motion condition and increased close to the stall point, but did not change significantly for the other motion conditions. Overall, pilot control behavior under enhanced hexapod motion was most similar to that under full aircraft motion. This indicates that motion cueing for stall recovery training on hexapod simulators might be improved by using the principles behind the enhanced hexapod motion configuration.

manual control↗
DOE OSTI2021

Auto-Curation of Seismic Event Data for Signal Denoising

Denoising contaminated seismic signals for later processing is a fundamental problem in seismic signals analysis. Neural network approaches have shown success denoising local signals when trained on short-time Fourier transform spectrograms. One challenge of this approach is the onerous process of hand-labeling event signals for training. By leveraging the SCALODEEP seismic event detector, we develop an automated set of techniques for labeling event data. Despite region specific challenges, training the neural network denoiser on machine curated events shows comparable performance to the neural network trained on hand curated events. We showcase our technique with two experiments, one using Utah regional data and one using regional data from the Korean peninsula.

58 GEOSCIENCES↗
NASA NTRS1989

Adjustment of sleep and the circadian temperature rhythm after flights across nine time zones

The adjustment of sleep-wake patterns and the circadian temperature rhythm was monitored in nine Royal Norwegian Airforce volunteers operating P-3 aircraft during a westward training deployment across nine time zones. Subjects recorded all sleep and nap times, rated nightly sleep quality, and completed personality inventories. Rectal temperature, heart rate, and wrist activity were continuously monitored. Adjustment was slower after the return eastward flight than after the outbound westward flight. The eastward flight produced slower readjustment of sleep timing to local time and greater interindividual variability in the patterns of adjustment of sleep and temperature. One subject apparently exhibited resynchronization by partition, with the temperature rhythm undergoing the reciprocal 15-h delay. In contrast, average heart rates during sleep were significantly elevated only after westward flight. Interindividual differences in adjustment of the temperature rhythm were correlated with some of the personality measures. Larger phase delays in the overall temperature waveform (as measured on the 5th day after westward flight) were exhibited by extraverts, and less consistently by evening types.

Gander, Philippa H.↗
DOE OSTI2024

Analysis and Benchmarking of feature reduction for classification under computational constraints

Abstract Machine learning is most often expensive in terms of computational and memory costs due to training with large volumes of data. Current computational limitations of many computing systems motivate us to investigate practical approaches, such as feature selection and reduction, to reduce the time and memory costs while not sacrificing the accuracy of classification algorithms. In this work, we carefully review, analyze, and identify the feature reduction methods that have low costs/overheads in terms of time and memory. Then, we evaluate the identified reduction methods in terms of their impact on the accuracy, precision, time, and memory costs of traditional classification algorithms. Specifically, we focus on the least resource intensive feature reduction methods that are available in Scikit-Learn library. Since our goal is to identify the best performing low-cost reduction methods, we do not consider complex expensive reduction algorithms in this study. In our evaluation, we find that at quadratic-scale feature reduction, the classification algorithms achieve the best trade-off among competitive performance metrics. Results show that the overall training times are reduced 61%, the model sizes are reduced 6×, and accuracy scores increase 25% compared to the baselines on average with quadratic scale reduction.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2010

Training to Facilitate Adaptation to Novel Sensory Environments

After spaceflight, the process of readapting to Earth s gravity causes locomotor dysfunction. We are developing a gait training countermeasure to facilitate adaptive responses in locomotor function. Our training system is comprised of a treadmill placed on a motion-base facing a virtual visual scene that provides an unstable walking surface combined with incongruent visual flow designed to train subjects to rapidly adapt their gait patterns to changes in the sensory environment. The goal of our present study was to determine if training improved both the locomotor and dual-tasking ability responses to a novel sensory environment and to quantify the retention of training. Subjects completed three, 30-minute training sessions during which they walked on the treadmill while receiving discordant support surface and visual input. Control subjects walked on the treadmill without any support surface or visual alterations. To determine the efficacy of training, all subjects were then tested using a novel visual flow and support surface movement not previously experienced during training. This test was performed 20 minutes, 1 week, and 1, 3, and 6 months after the final training session. Stride frequency and auditory reaction time were collected as measures of postural stability and cognitive effort, respectively. Subjects who received training showed less alteration in stride frequency and auditory reaction time compared to controls. Trained subjects maintained their level of performance over 6 months. We conclude that, with training, individuals became more proficient at walking in novel discordant sensorimotor conditions and were able to devote more attention to competing tasks.

Bloomberg, J. J.↗
DOE OSTI2026

F-Hash: Feature-Based Hash Design for Time-Varying Volume Visualization via Multi-Resolution Tesseract Encoding

Interactive time-varying volume visualization is challenging due to its complex spatiotemporal features and sheer size of the dataset. Recent works transform the original discrete time-varying volumetric data into continuous Implicit Neural Representations (INR) to address the issues of compression, rendering, and super-resolution in both spatial and temporal domains. However, training the INR takes a long time to converge, especially when handling large-scale time-varying volumetric datasets. In this work, we proposed F-Hash, a novel feature-based multi-resolution Tesseract encoding architecture to greatly enhance the convergence speed compared with existing input encoding methods for modeling time-varying volumetric data. The proposed design incorporates multi-level collision-free hash functions that map dynamic 4D multi-resolution embedding grids without bucket waste, achieving high encoding capacity with compact encoding parameters. Our encoding method is agnostic to time-varying feature detection methods, making it a unified encoding solution for feature tracking and evolution visualization. Experiments show the F-Hash achieves state-of-the-art convergence speed in training various time-varying volumetric datasets for diverse features. We also proposed an adaptive ray marching algorithm to optimize the sample streaming for faster rendering of the time-varying neural representation.

deep learning↗
DOE OSTI2024

Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements

In H-mode tokamak plasmas, the plasma is sometimes ejected beyond the edge transport barrier. These events are known as edge localized modes (ELMs). ELMs cause a loss of energy and damage the vessel walls. Understanding the physics of ELMs, and by extension, how to detect and mitigate them, is an important challenge. In this paper, we focus on two diagnostic methods—deuterium-alpha (D α ) spectroscopy and Doppler backscattering (DBS). The former detects ELMs by measuring Balmer alpha emission, while the latter uses microwave radiation to probe the plasma. DBS has the advantages of having a higher temporal resolution and robustness to damage. These advantages of DBS diagnostic may be beneficial for future operational tokamaks, and thus, data processing techniques for DBS should be developed in preparation. In sight of this, we explore the training of neural networks to detect ELMs from DBS data, using D α data as the ground truth. With shots found in the DIII-D database, the model is trained to classify each time step based on the occurrence of an ELM event. The results are promising. When tested on shots similar to those used for training, the model is capable of consistently achieving a high f1-score of 0.93. Furthermore, this score is a performance metric for imbalanced datasets that ranges between 0 and 1. We evaluate the performance of our neural network on a variety of ELMs in different high confinement regimes (grassy ELM, RMP mitigated, and wide-pedestal), finding broad applicability. Beyond ELMs, our work demonstrates the wider feasibility of applying neural networks to data from DBS diagnostic.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRS2014

High Performance EVA Glove Collaboration: Glove Injury Data Mining Effort

Human hands play a significant role during extravehicular activity (EVA) missions and Neutral Buoyancy Lab (NBL) training events, as they are needed for translating and performing tasks in the weightless environment. It is because of this high frequency usage that hand- and arm-related injuries and discomfort are known to occur during training in the NBL and while conducting EVAs. Hand-related injuries and discomforts have been occurring to crewmembers since the days of Apollo. While there have been numerous engineering changes to the glove design, hand-related issues still persist. The primary objectives of this study are therefore to: 1) document all known EVA glove-related injuries and the circumstances of these incidents, 2) determine likely risk factors, and 3) recommend ergonomic mitigations or design strategies that can be implemented in the current and future glove designs. METHODS: The investigator team conducted an initial set of literature reviews, data mining of Lifetime Surveillance of Astronaut Health (LSAH) databases, and data distribution analyses to understand the ergonomic issues related to glove-related injuries and discomforts. The investigation focused on the injuries and discomforts of U.S. crewmembers who had worn pressurized suits and experienced glove-related incidents during the 1980 to 2010 time frame, either during training or on-orbit EVA. In addition to data mining of the LSAH database, the other objective of the study was to find complimentary sources of information such as training experience, EVA experience, suit-related sizing data, and hand-arm anthropometric data to be tied to the injury data from LSAH. RESULTS: Past studies indicated that the hand was the most frequently injured part of the body during both EVA and NBL training. This study effort thus focused primarily on crew training data in the NBL between 2002 and 2010. Of the 87 recorded training incidents, 19 occurred to women and 68 to men. While crew ages ranged from thirties to fifties, the age category most affected was in the forties range. Incident rate calculations (incidents per 100 training runs) revealed that the 2002, 2003, and 2004 time periods registered the highest reported incident rate levels (3.4, 6.1, and 4.1 respectively) when compared to the following years (all ≤ 1.0). In addition to general hand-arm discomfort being the highest reported result from training, specific types of hand injuries or symptoms included erythema, fingernail delamination, abrasions, muscle soreness/fatigue, paresthesia, bruising, blanching, and edema. Specific body locations most affected by hand injuries included the metacarpophalangeal joints, fingernails, finger crotches, fingers in general, interphalangeal joints, and fingertips. Causes of injuries reported in the LSAH data were primarily attributed to the forces that the gloved hands were exposed to due to hand intensive tasks and/or poor glove sizing. DISCUSSION: Although the age data indicate that most injuries are reported by male crewmembers in their forties, that is also the dominant gender and age range of most EVA crew therefore it is not an unexpected finding. Age and gender analysis will continue as more details on the uninjured population is accrued. While there is a reasonable mechanism to link training quantity to injury, the results were inconsistent and point to the need for a consistent method of suit-related injury screening and documentation. For instance, the high-incident rate levels for the years 2002 to 2004 could be attributed to a comprehensive medical review of crewmembers post-NBL EVA training that occurred from July 19, 2002 to January 16, 2004. Furthermore, there could have been increased awareness from an investigation at the NBL. These investigations may have temporarily increased the fidelity of reported injuries and discomforts during these dates as compared to surrounding years, when injury signs and symptom were no longer actively being investigated but rather voluntarily reported. Data mining for possible mechanistic factors continues and includes more detailed training timelines, hand anthropometry, and suit sizing information. The limited published data looking at hand-arm anthropometry correlated hand-anthropometry metrics with injuries stemming from glove design and operation. Future work will include further evaluation of body sizing and fit in relation to hand injury incidents.

Reid, C. R.↗
NASA NTRS1987

Martian surface physical properties to be derived by radar altimeter on the Mars observer spacecraft

The potential is described of a candidate Mars Observer altimeter for determining dielectric properties of Mars regolith. It is pointed out that it is straightforward to use the time between altimeter pulse trains for passive radiometry (hence dielectric properties) and roughness can be derived. Given the mission plan the whole surface can be mapped at least three times, yielding data on seasonal variability.

Garvin, J. B.↗