SEARCH · Engineering Papers

Results for “validation dataset”

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 1,603 records · Page 89

All sources NASA NTRS DOE OSTI

DOE OSTI2024

Acoustic-based monitoring and machine learning of component status for microreactor applications

This report provides a description and assessment of recent efforts to couple acoustic-based experimental measurements and characterization with machine learning models in order to enhance structural health monitoring capabilities for nuclear microreactors. With resilient embedded sensors in development by others supported by programs funded by the US Department of Energy’s Office of Nuclear Energy, the work described herein builds upon ongoing efforts to improve non-destructive testing technology that relates measured acoustic signatures to component stresses and/or structural defects, using a combination of new experimental measurements and machine learning architectures. The experimental procedure remained similar to that developed for the previous year’s demonstration of damage detection by the authors, with the same damaged sample tested under similar applied stress conditions. Notably, a new mounting fixture was designed and implemented to improve measurement consistency and a more sophisticated laser Doppler vibrometer was employed to make high-fidelity vibration measurements. Two nominally identical sets of training data were collected for each experimental setup to better understand the repeatability of the experiment and to better test the generality of trained neural network models. Additionally, we obtained new high-quality 3D mode shapes of the damaged test article at various stress and excitation levels, providing greater insights into the physical response of the sample during testing. Previously, we demonstrated that a machine learning model based on a convolutional neural network can predict structural details of an artificially introduced interface (intact, rough cut, smooth cut), and the applied torque level. In this study, we have transitioned to graph-based neural network architectures to better develop and test a flexible framework that is more suitable to being transferred away from controlled benchtop experiments and into more applied settings where less-structured data inputs may be expected. In general, performance testing of a graph neural network on frequency-domain representations of the data indicates strong and consistent identification of test conditions for datasets recorded on damaged components. With goals of predicting damage location and other changing experimental conditions using limited datasets, predictive models using a graph neural network architecture correctly predicted the applied torque level with an accuracy of 85% using only a single measurement point and predicted within one torque level in 95% of test windows. Predictions of damage location had limited success due to the symmetry and minimal number of the damage scenarios presented during model training. Results were ambiguous as to whether the model could detect the location of the artificial damage, or if it was instead learning the location of a given measurement point on the part and subsequently detecting which points were closest to the location of the damage. This finding will be factored into upcoming planned work on damaged graphite components, where new experimental tests with a larger number and variety of damage scenarios are expected to provide improved validation of recent developments in monitoring methodology.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

NASA NTRS2022

Cloud Mask Intercomparison eXercise (CMIX): An evaluation of cloud masking algorithms for Landsat 8 and Sentinel-2

Cloud cover is a major limiting factor in exploiting time-series data acquired by optical spaceborne remote sensing sensors. Multiple methods have been developed to address the problem of cloud detection in satellite imagery and a number of cloud masking algorithms have been developed for optical sensors but very few studies have carried out quantitative intercomparison of state-of-the-art methods in this domain. This paper summarizes results of the first Cloud Masking Intercomparison eXercise (CMIX) conducted within the Committee Earth Observation Satellites (CEOS) Working Group on Calibration & Validation (WGCV). CEOS is the forum for space agency coordination and cooperation on Earth observations, with activities organized under working groups. CMIX, as one such activity, is an international collaborative effort aimed at intercomparing cloud detection algorithms for moderate-spatial resolution (10–30 m) spaceborne optical sensors. The focus of CMIX is on open and free imagery acquired by the Landsat 8 (NASA/USGS) and Sentinel-2 (ESA) missions. Ten algorithms developed by nine teams from fourteen different organizations representing universities, research centers and industry, as well as space agencies (CNES, ESA, DLR, and NASA), are evaluated within the CMIX. Those algorithms vary in their approach and concepts utilized which were based on various spectral properties, spatial and temporal features, as well as machine learning methods. Algorithm outputs are evaluated against existing reference cloud mask datasets. Those datasets vary in sampling methods, geographical distribution, sample unit (points, polygons, full image labels), and generation approaches (experts, machine learning, sky images). Overall, the performance of algorithms varied depending on the reference dataset, which can be attributed to differences in how the reference datasets were produced. The algorithms were in good agreement for thick cloud detection, which were opaque and had lower uncertainties in their identification, in contrast to thin/semi-transparent clouds detection. Not only did CMIX allow identification of strengths and weaknesses of existing algorithms and potential areas of improvements, but also the problems associated with the existing reference datasets. The paper concludes with recommendations on generating new reference datasets, metrics, and an analysis framework to be further exploited and additional input datasets to be considered by future CMIX activities.

Sergii Skakun↗

DOE OSTI2025

Edge ML for CAN bus intrusion detection in AVs

Autonomous Vehicles (AVs) are revolutionizing transportation, but their reliance on interconnected cyber-physical systems exposes them to unprecedented cybersecurity risks. This study addresses the critical challenge of detecting real-time cyber intrusions in self-driving vehicles by leveraging a dataset from the Udacity self-driving car project. We simulate four high-impact attack vectors, Denial of Service (DoS), spoofing, replay, and fuzzy attacks, by injecting noise into spatial features (e.g., bounding box coordinates) to replicate adversarial scenarios. We develop and evaluate two lightweight neural network architectures (NN-1 and NN-2) alongside a logistic regression baseline (LG-1) for intrusion detection. The models achieve exceptional performance, with NN-2 attaining an AUC score of 93.15% and 93.15% accuracy, demonstrating their suitability for edge deployment in AV environments. Through explainable AI techniques, we uncover unique forensic fingerprints of each attack type, such as spatial corruption in fuzzy attacks and temporal anomalies in replay attacks, offering actionable insights for feature engineering and proactive defense. Visual analytics, including confusion matrices, ROC curves, and feature importance plots, validate the models' robustness and interpretability. This research sets a new benchmark for AV cybersecurity, delivering a scalable, field-ready toolkit for Original Equipment Manufacturers (OEMs) and policymakers. By aligning intrusion fingerprints with SAE J3061 automotive security standards, we provide a pathway for integrating machine learning into safety-critical AV systems. Our findings underscore the urgent need for security-by-design AI, ensuring that AVs not only drive autonomously but also defend autonomously. This work bridges the gap between theoretical cybersecurity and life-preserving engineering, offering a leap toward safer, more secure autonomous transportation.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS2017

Risk-Significant Adverse Condition Awareness Strengthens Assurance of Fault Management Systems

As spaceflight systems increase in complexity, Fault Management (FM) systems are ranked high in risk-based assessment of software criticality, emphasizing the importance of establishing highly competent domain expertise to provide assurance. Adverse conditions (ACs) and specific vulnerabilities encountered by safety- and mission-critical software systems have been identified through efforts to reduce the risk posture of software-intensive NASA missions. Acknowledgement of potential off-nominal conditions and analysis to determine software system resiliency are important aspects of hazard analysis and FM. A key component of assuring FM is an assessment of how well software addresses susceptibility to failure through consideration of ACs. Focus on significant risk predicted through experienced analysis conducted at the NASA Independent Verification Validation (IVV) Program enables the scoping of effective assurance strategies with regard to overall asset protection of complex spaceflight as well as ground systems. Research efforts sponsored by NASA's Office of Safety and Mission Assurance defined terminology, categorized data fields, and designed a baseline repository that centralizes and compiles a comprehensive listing of ACs and correlated data relevant across many NASA missions. This prototype tool helps projects improve analysis by tracking ACs and allowing queries based on project, mission type, domaincomponent, causal fault, and other key characteristics. Vulnerability in off-nominal situations, architectural design weaknesses, and unexpected or undesirable system behaviors in reaction to faults are curtailed with the awareness of ACs and risk-significant scenarios modeled for analysts through this database. Integration within the Enterprise Architecture at NASA IVV enables interfacing with other tools and datasets, technical support, and accessibility across the Agency. This paper discusses the development of an improved workflow process utilizing this database for adaptive, risk-informed FM assurance that critical software systems will safely and securely protect against faults and respond to ACs in order to achieve successful missions.

Fault management↗

NASA NTRS2017

GC13I-0860: An Assessment of Surface Water Detection Methods for the Tahoua Region, Niger

The recent release of several global surface water datasets derived from remotely sensed data has allowed for unprecedented analysis of the earth's hydrologic processes at a global scale. However, some of these datasets fail to identify important sources of surface water, especially small ponds, in the Sahel, an arid region of Africa that forms a border zone between the Sahara Desert to the north, and the savannah to the south. These ponds may seem insignificant in the context of wider, global-scale hydrologic processes, but smaller sources of water are important for local and regional hydrologic assessments. Particularly, these smaller water bodies are significant sources of hydration and irrigation for nomadic pastoralists and smallholder farmers throughout the Sahel. For this study, several methods of identifying surface water from Landsat 8 OLI, Sentinel 1 SAR, Sentinel 2 MSI, and Planet Dove data were compared to determine the most effective means of delineating these features in the Tahoua Region of Niger. The Automated Water Extraction Index (AWEInsh) had the best performance when validated against very high resolution Digital Globe imagery, with an overall accuracy of 98.6%. This study reiterates the importance of region-specific algorithms and suggests that the AWEInsh method may be the best for delineating surface water in the Sahelian ecozone, likely due to the nature of the exposed geology and lack of dense green vegetation.

Herndon, Kelsey E.↗

NASA NTRS2017

Meteorological Inventory of Rain-On-Snow Events in the Canadian Arctic Archipelago and Satellite Detection Assessment Using Passive Microwave Data

The spatial and temporal distributions of rain-on-snow (ROS) events across the Canadian Arctic Archipelago (CAA) remain poorly understood owing to their sporadic nature in time and space. This situation motivated the development of remote sensing detection algorithms. This paper uses a large meteorological dataset across the CAA to adapt an existing ROS-detection algorithm developed in a previous study by our group. Results highlight the spatial distribution and evolution of ROS occurrences reported since 1985 at 14 weather stations across the CAA. Results show that >600 ROS events were inventoried since 1985, for which >70% were classified as pure rain (liquid form) and 30% as mixed precipitation (solid/liquid). Of the pure rain events, 75% occurred during spring, 14% during fall, 8% during summer and <1% during winter. Such events can have significant impacts on ungulate grazing conditions through the creation of ice layers, causing serious problems for caribou calf survival, especially during the migration period. This paper introduces an adaptation for larger scale Arctic application of a detection algorithm (sensitivity analysis on the detection threshold) with an error of ~5%. The validation, however, remains limited due to a short study period and limited number of sites.

Canadian Arctic Archipelago (CAA)↗

DOE OSTI2020

NEON AOP foliar trait maps, maps of model uncertainty estimates, and conifer map, East River, CO 2018

This data package contains mapped trait estimates and their uncertainties, and conifer map, for the National Ecological Observatory Network's Airborne Observation Platform survey data acquired over the Upper East River, Colorado in 2018. For full details, please see associated reference. in brief, trait models were developed independently for needle and non-needle leaf species using partial least squares regression (PLSR) using ground data from additional datasets: doi:10.15485/1618130, doi:10.15485/1618132, and doi:10.15485/1631278, merged with extracted spectral data from doi:10.15485/1618131. We separated vegetated pixels into needle and non-needle classes in order to generate a classification map based on the spectral differences between these leaf types (conifer.tif). We trained a deep learning model with custom architecture, detailed in Chadwick et al. In Press. The model performed with 0.998 true positive rate and 0.982 true negative rate, with ‘positives’ being non-needle identification. We then utilized PLSR to generate models of foliar traits for each leaf type. So that we could also map uncertainty in these predictions, we generated ten different models for needle and non-needle leaf species using different testing holdout sets of discrete sites. Each of these models was developed with a 100-fold cross validation procedure that utilized a 70% training set and 30% validation set with each fold, and then assessed based on the 10% of testing sites that were not included in that model’s development. The mean predicted value across the 10 models is used for the trait estimate in each pixel across the study area. The models are applied according to the leaf type designation in the conifer.tif map. The errors are the standard deviation across the 10 different models developed, with high error suggesting instability in model prediction and areas where values may not be reliable for ecological inference. These maps are only applied to areas with a NDVI > 0.5 to exclude non-vegetated areas. Shade masks could be applied to these data (doi:10.15485/1618131), but have not been for this data package. These data are also available on Google Earth Engine: https://code.earthengine.google.com/?asset=users/kdc/ER_NEON

54 ENVIRONMENTAL SCIENCES↗

NASA NTRSDate not supplied

Quantifying radiation quality for space relevant radiation types: Fitting excess risk models to three combined HZE-irradiated mouse datasets

Radiation health risks are predominantly derived from low linear energy transfer (LET) terrestrial exposures; however, space radiation includes exposure to high-LET and high-charge, high-energy (HZE) particles. Accurately quantifying the differences in radiation quality between the space and terrestrial radiation environments is important for assessing and predicting health risks for astronauts. Weil et al. 2009 and 2014 used two different inbred mouse strains to study differences in hepatocellular carcinoma (HCC) tumorigenesis after exposures to low- and high- LET radiation. More recently, Edmundson et al. 2020 provided valuable new tumor data in outbred mice that were exposed to low- and high-LET radiation. The present study aims to rigorously investigate a relative biological effectiveness (RBE) factor by leveraging the HCC tumor data from Weil et al. 2009, Weil et al. 2014, and Edmundson et al. 2020. The three experiments were similarly designed, allowing the raw data to be combined into a pooled dataset to estimate excess relative risk (ERR) and excess absolute risk (EAR) models using Bayesian Poisson regression. These effect estimates from the pooled data provide greater power to calculate a data driven RBE. Extensive sensitivity analyses test the robustness of RBE estimates to various model assumptions. The following questions will be explored through the sensitivity analyses: • Is the shape of the dose response different for low-LET radiation and HZE radiation, indicating that RBE is a function of dose? • Does attained age modify the effect estimates differently for low-LET radiation and HZE radiation, indicating RBE is a function of attained age? • Are the effect estimates and RBE estimates different for inbred mouse strains and outbred mouse strains? • Do assumptions about differences in ERR models and EAR models change the estimated RBE? Additional studies would be needed to validate the findings from these exploratory analyses.

Lori J. Chappell↗

DOE OSTI2021

Detecting forest response to droughts with global observations of vegetation water content

Droughts in a warming climate have become more common and more extreme, making understanding forest responses to water stress increasingly pressing. Analysis of water stress in trees has long focused on water potential in xylem and leaves, which influences stomatal closure and water flow through the soil-plant-atmosphere continuum. At the same time, changes of vegetation water content (VWC) are linked to a range of tree responses, including fluxes of water and carbon, mortality, flammability, and more. Unlike water potential, which requires demanding in situ measurements, VWC can be retrieved from remote sensing measurements, particularly at microwave frequencies using radar and radiometry. Here, we highlight key frontiers through which VWC has the potential to significantly increase our understanding of forest responses to water stress. To validate remote sensing observations of VWC at landscape scale and to better relate them to data assimilation model parameters, we introduce an ecosystem-scale analogue of the pressure-volume curve, the non-linear relationship between 44 average leaf or branch water potential and water content commonly used in plant hydraulics. The 45 sources of variability in these ecosystem-scale pressure-volume curves and their relationship to 46 forest response to water stress are discussed. We further show to what extent diel, seasonal, and 47 decadal dynamics of VWC reflect variations in different processes relating the tree response to 48 water stress. VWC can also be used for inferring belowground conditions – which are difficult to 49 impossible to observe directly. Lastly, we discuss how a dedicated geostationary spaceborne 50 observational system for VWC, when combined with existing datasets, can capture diel and 51 seasonal water dynamics to advance the science and applications of global forest vulnerability to 52 future droughts.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS2016

Mach Cutoff Analysis and Results from NASA's Farfield Investigation of No-Boom Thresholds

In support of the ongoing effort by the National Aeronautics and Space Administration (NASA) to bring supersonic commercial travel to the public, the NASA Armstrong Flight Research Center and the NASA Langley Research Center, in partnership with other industry organizations and academia, conducted a flight research experiment to analyze acoustic propagation in the Mach cutoff shadow zone. The effort was conducted in the fall of 2012 and named the Farfield Investigation of No-boom Thresholds (FaINT). The test helped to build a dataset that will go toward further understanding of the unique acoustic propagation characteristics below Mach cutoff altitude. FaINT was able to correlate sonic boom noise levels measured below cutoff altitude with precise airplane flight conditions, potentially increasing the accuracy over previous studies. A NASA F-18B airplane made supersonic passes such that its Mach cutoff caustic would be at varying distances above a linear 60-microphone, 7375-ft (2247.9 m) long array. A TG-14 motor glider equipped with a microphone on its wing-tip also attempted to capture the same sonic boom waves above ground, but below the Mach cutoff altitude. This paper identified an appropriate metric for sonic boom waveforms in the Mach cutoff shadow zone called Perceived Sound Exposure Level; derived an empirical relationship between Mach cutoff flight conditions and noise levels in the shadow zone; validated a safe cutoff altitude theory presented by previous studies; analyzed the sensitivity of flight below Mach cutoff to unsteady atmospheric conditions and realistic aircraft perturbations; and demonstrated the ability to record sonic boom measurements over 5000 ft (1524.0 m) above ground level, but below Mach cutoff altitude.

flight research↗

NASA NTRS2017

Risk-Significant Adverse Condition Awareness Strengthens Assurance of Fault Management Systems

As spaceflight systems increase in complexity, Fault Management (FM) systems are ranked high in risk-based assessment of software criticality, emphasizing the importance of establishing highly competent domain expertise to provide assurance. Adverse conditions (ACs) and specific vulnerabilities encountered by safety- and mission-critical software systems have been identified through efforts to reduce the risk posture of software-intensive NASA missions. Acknowledgement of potential off-nominal conditions and analysis to determine software system resiliency are important aspects of hazard analysis and FM. A key component of assuring FM is an assessment of how well software addresses susceptibility to failure through consideration of ACs. Focus on significant risk predicted through experienced analysis conducted at the NASA Independent Verification & Validation (IV&V) Program enables the scoping of effective assurance strategies with regard to overall asset protection of complex spaceflight as well as ground systems. Research efforts sponsored by NASAs Office of Safety and Mission Assurance (OSMA) defined terminology, categorized data fields, and designed a baseline repository that centralizes and compiles a comprehensive listing of ACs and correlated data relevant across many NASA missions. This prototype tool helps projects improve analysis by tracking ACs and allowing queries based on project, mission type, domain/component, causal fault, and other key characteristics. Vulnerability in off-nominal situations, architectural design weaknesses, and unexpected or undesirable system behaviors in reaction to faults are curtailed with the awareness of ACs and risk-significant scenarios modeled for analysts through this database. Integration within the Enterprise Architecture at NASA IV&V enables interfacing with other tools and datasets, technical support, and accessibility across the Agency. This paper discusses the development of an improved workflow process utilizing this database for adaptive, risk-informed FM assurance that critical software systems will safely and securely protect against faults and respond to ACs in order to achieve successful missions.

IV&V↗

NASA NTRSDate not supplied

Acquisition of and Access to Research Omics Data

Omics data are essential for understanding the myriad and complex effects of space environments on humans. To assure maximum benefit from these kinds of data, the NASA Human Research Program Data Management Plan stipulates that human omics data should be archived within and accessed through the NASA Life Sciences Portal (NLSP). The NLSP has the capability to acquire and provision access to omics (and other kinds of) research results for individual and ad-hoc groups of subjects at the direction of institutional review boards, or other authorizing bodies or individuals, per institutional, program and investigation-specific policies and procedures. However, because some single-subject omics data, like CT scans and other kinds of large, complex biomedical data, could be used to identify heretofore unknown risks to the subject’s health, or, in certain cases, be used to identify a subject, NASA Policy Directive 7170.1 describes various policies regarding the management of and access to “research genetic testing” data, which includes many kinds of omics data. For example, NPD 7170.1 prohibits access to human research genetic data by NASA personnel who make employment decisions for the subjects from whom the data were obtained. To meet the objective of acquiring research omics data for NLSP in compliance with the policies in NPD 7170.1 and other applicable NASA policies, we designed NOMADS (the NLSP Omics Multimodal Acquisition of Data System), a new component that supports the transfer of large research data files, including research genetic testing data, using one of several different transfer mechanisms. The choice of mechanism is made by the submitter of the data, with guiding information from the system, and is likely to often be determined in large part by the nature and source location of the data. For example, for small files where the source data files are not already stored in a cloud storage system, users are likely to prefer to transfer their data to the NLSP via a web browser. Conversely, for large sets of files already organized and stored in a cloud storage system, users may opt for NOMAD’s cloud-to-cloud transfer method. All omics datasets targeted for the NASA Life Sciences Data Archive must pass a variety of quality checks to ensure data integrity and adherence to the standards defined by the LSDA Data Submission Guidelines (DSG) (see https://nlsp.nasa.gov/explore/lsdahome/datasubmit). These include requirements that data are consistent with open standards established by the omics community. Non-compliant data will not be accepted however archivists are available to advise submitters on how to revise data submissions and re-submit until compliance is achieved. Following compliance with the LSDA DSG, omics data next undergo a variety of additional quality checks to ensure the data meet omics community standards. Domain specific Omics data quality control tools and techniques are continually evolving and linked to the advancements in omics assays utilized and thus, the tools and techniques utilized by the LSDA for data quality control and validation will need to be sustained accordingly. All human omics data will be access controlled according to the policies described above, and requiring IRB approval for any additional access grants once the data are acquired (including access for analysis using the NLSP workspace tools).

Omics↗

NASA NTRS2024

Acquisition of and Access to Research Omics Data

Omics↗

DOE OSTI2022

End-Use Load Profiles for the U.S. Building Stock: Methodology and Results of Model Calibration, Validation, and Uncertainty Quantification

The United States is embarking on an ambitious transition to a 100% clean energy economy by 2050, which will require improving the flexibility of electric grids. One way to achieve grid flexibility is to shed or shift demand to align with changing grid needs. To facilitate this, it is critical to understand how and when energy is used. High- quality end-use load profiles (EULPs) provide this information, and can help cities, states, and utilities understand the time-sensitive value of energy efficiency, demand response, and distributed energy resources. Publicly available EULPs have traditionally had limited application because of age and incomplete geographic representation (Frick, Eckman, and Goldman 2017; Frick 2019). To help fill this gap, the U.S. Department of Energy (DOE) funded a three-year project - End-Use Load Profiles for the U.S. Building Stock - that culminated in the release of a publicly available dataset1 of simulated EULPs representing residential and commercial buildings across the contiguous United States. The motivation for this work is further detailed in a November 2019 report: Market Needs, Use Cases, and Data Gaps (Mims Frick et al. 2019). This Methodology and Results report provides detailed descriptions of how the dataset was developed, intended for an audience of dataset and model users interested in the technical details. These details include descriptions of all of the model improvements made for calibration and the final comparisons to empirical data sources. A companion report, End-Use Load Profiles for the U.S. Building Stock: Applications and Opportunities, will be published subsequently and will describe example applications and considerations for using the dataset, intended for an audience of general dataset users.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

NASA NTRSDate not supplied

Algorithmic Classification of Raman Spectra Biosignatures: Improving Life Detection Confidence

“Agnostic” biosignatures – indicators of life (or the absence of life), independent of a particular biochemistry – are increasingly considered a high standard for life detection. The Ladder of Life Detection (2018) called for investigating how combinations of independent and different potential biosignatures affect confidence. To address this gap, statistical classification of elemental abundances, isotopic fractionation, and reflectance spectroscopy (VNIR) has been implemented. Raman spectroscopy, highly desirable due to its wide availability, has the potential to improve this predictive power. This work implemented biosignature classification algorithms on Raman data alone, in preparation for combination with the other data types. Raman spectroscopy data was collected from published databases and papers as part of a manually curated dataset of “indicative” and “non-indicative of life” samples. These currently include 61 non-indicative samples (meteorites, magnetite); 3 indicative living samples (bacteria); 20 indicative non-living samples (chalk, bone); and 12 indicative mixed (with non-indicative material) samples (soil, microbial mats). Laboratory work is ongoing to characterize additional samples, particularly a greater breadth of mixed systems. Spectra were interpolated, filtered with the Savitzsky-Golay filter, and de-noised. For a preliminary examination, agnostic features were manually extracted including mean intensity, number of peaks, and mean peak width. Different peak prominences and filtering polynomials were used to refine features. Classification algorithms were implemented: k-nearest neighbors (KNN), logistic regression (LR), linear support vector machines (SVM), random forest (RF), Gaussian naïve bayes (GNB). Lastly, Monte Carlo simulations on 1,000 50%-train-test-splits were used to validate classification performance and feature significance. The preliminary feature set achieved its highest AUC of 0.52 with LR, with no strongly discriminatory features. Work to improve feature extraction, such as through deep learning with back propagation, is planned. In future work, the Raman data will be combined with the other data types, and potentially new data types such as enantiomeric excess. This project was partially supported through the NASA Ames Project EXcellence (APEX) incubator program.

Astrobiology↗

NASA NTRSDate not supplied

From Simulation to Reality With Random Noise

The challenging environment of autonomous vehicle (AV) navigation necessitates certain functions be performed by deep neural networks. Optimizing these models involves collecting vast quantities of domain-specific training data and ensuring that the dataset is representative of expected conditions. High-fidelity simulation plays a vital role in making this process feasible, allowing a wide range of scenarios to be explored at low cost. However, learning from simulation introduces subtle biases into models, which can degrade real-world performance in unpredictable ways. This effect can be mitigated with learning schemes specialized to bridge distributional shifts (transfer learning). Given the complex nature of these methods, the underlying models, and their environments, meaningfully evaluating performance is notstraight forward. Many unrelated factors can effect an improvement in generalization accuracy, but a full ablation analysis is often difficult. To tease out signal from noise, it is necessary to understand how transfer learning performance is affected by noise itself. The goals of this paper are (i) to establish a domain randomization baseline for a simple classification transfer learning task and (ii) to validate the RRAV testbed as a platform for further research in sim-to-real learning. We generate imagery from a simulation of NASA Ames Research Center and train a small convolutional neural network (ConvNet) to classify position relative to a centerline. Further models are trained with different types of noise progressively added to the data. The models are deployed aboard the on-site test vehicle to test real-world performance. In our experiments, we find that such naive domain randomization raises sim-to-real accuracy from 64% to 79%, while training directly on real data yields an 89% accuracy ceiling. These results suggest that the isolated mechanism of domain randomization can significantly improve generalization.

simulation↗

DOE OSTI2025

An analysis of parameter compression and Full-Modeling techniques with Velocileptors for DESI 2024 and beyond

In anticipation of forthcoming data releases of current and future spectroscopic surveys, we present the validation tests and analysis of systematic effects within velocileptors modeling pipeline when fitting mock data from the AbacusSummit N-body simulations. We compare the constraints obtained from parameter compression methods to the direct fitting (Full-Modeling) approaches of modeling the galaxy power spectra, and show that the ShapeFit extension to the traditional template method is consistent with the Full-Modeling method within the standard ΛCDM parameter space. We show the dependence on scale cuts when fitting the different redshift bins using the ShapeFit and Full-Modeling methods. We test the ability to jointly fit data from multiple redshift bins as well as joint analysis of the pre-reconstruction power spectrum with the post-reconstruction BAO correlation function signal. We further demonstrate the behavior of the model when opening up the parameter space beyond ΛCDM and also when combining likelihoods with external datasets, namely the Planck CMB priors. Finally, we describe different parametrization options for the galaxy bias, counterterm, and stochastic parameters, and employ the halo model in order to physically motivate suitable priors that are necessary to ensure the stability of the perturbation theory.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2026

Microstructure Validation of Graph Theory Model-Derived Cooling Rates in the Wire Arc Additive Manufacturing of ER70S-6 Steel

Wire arc additive manufacturing (WAAM) enables high-rate fabrication of large metallic components, but spatial variations in thermal history can lead to microstructural heterogeneity that requires efficient process models to evaluate. This study evaluates whether cooling rates extracted from a graph theory model (GTM)-based thermal simulation are consistent with the microstructural evolution observed in an ER70S-6 WAAM wall. Thermal histories from the model were analyzed at selected build heights, and cooling rates were extracted from the final thermal excursion through the austenite phase field. Microstructures at corresponding locations were characterized using electron backscatter diffraction (EBSD) to quantify grain size distributions, and pearlite interlamellar spacing was used as an additional indicator of cooling behavior. The modeled cooling rates were highest near the substrate and generally decreased with build height, consistent with the observed reduction in the fine grain fraction and the progressive shift in the grain size distribution as build height increased. Pearlite spacing trends also supported the modeled cooling rate variation. These results indicate that GTM-derived thermal histories can be post-processed into metallurgically meaningful cooling rate estimates for WAAM steel builds and linked to dataset specific empirical grain size distribution relationships for process–thermal history–microstructure assessment.

36 MATERIALS SCIENCE↗