SEARCH · Engineering Papers
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.
NASA Electronic Parts and Packaging (NEPP) is developing a process to be employed in critical applications. The framework assesses levels of trust and assurance in microelectronic systems. The process is being created with participation from a variety of organizations. We present a synopsis of the framework that includes contributions from The Aerospace Corporation.
Here, this work proposes a framework for large-scale stochastic derivative-free optimization (DFO) by introducing STARS, a trust-region method based on iterative minimization in random subspaces. This framework is both an algorithmic and theoretical extension of a random subspace derivative-free optimization (RSDFO) framework, and an algorithm for stochastic optimization with random models (STORM). Moreover, like RSDFO, STARS achieves scalability by minimizing interpolation models that approximate the objective in low-dimensional affine subspaces, thus significantly reducing per-iteration costs in terms of function evaluations and yielding strong performance on largescale stochastic DFO problems. The user-determined dimension of these subspaces, when the latter are defined, for example, by the columns of so-called Johnson-Lindenstrauss transforms, turns out to be independent of the dimension of the problem. For convergence purposes, inspired by the analyses of RSDFO and STORM, both a particular quality of the subspace and the accuracies of random function estimates and models are required to hold with sufficiently high, but fixed, probabilities. Using martingale theory under the latter assumptions, an almost sure global convergence of STARS to a first-order stationary point is shown, and the expected number of iterations required to reach a desired first-order accuracy is proved to be similar to that of STORM and other stochastic DFO algorithms, up to constants.
SPECTRE-ML (Spectral Peak Enhancement by Combining Trusted Response Elements via Machine Learning) is a machine learning based program for finding optimal clusters of radiation detector segments (i.e., pixels or voxels) in order to improve spectral performance. It provides facilities for pre-processing and analyzing training datasets, running ML algorithms, and evaluating and visualizing outputs. SPECTRE-ML outperforms simpler ad-hoc segmentation methods such as uniform depth clusters, learning detector performance trends such as dead layers, edge effects, and gain shifts. Although extensible to arbitrary highly-segmented spectroscopic radiation detectors, SPECTRE-ML currently focuses on improving spectral performance in highly-segmented CdZnTe (CZT) detectors for International Atomic Energy Agency (IAEA) non-destructive assay (NDA) safeguards tasks.
The TRUST testbed was designed and fabricated to conduct vibration testing and validate the behavior of hyper-elastic foams in a pre-loaded condition. The main goals of this testbed design were to ensure symmetry, proper instrumentation alignment and mounting, simplified geometry to make the data analysis validation process more straightforward, and to apply and monitor the applied pre-load. Based off of the test parameters, a shaker system was purchased. The baseplate was designed to mount directly to the bolt holes provided on the top of the shaker table. It was determined that the specimens will be cylindrical and vary in thickness from 2-10mm. The baseplate riser, center-mass, and testbed cap were designed to be cylindrical to make alignment about the center axis more straightforward. It was determined to fabricate the parts previously mentioned using Aluminum 6061-T6 due to its stiffness in order to reduce the likelihood of the stiff material impacting the data being collected for the foam specimens. The design concept used to apply the pre-load to the testbed was a through bolt/nut configuration. A through hole was included in each piece of the testbed, specimens, and instrumentation large enough to allow for 1/8” clearance between the through bolt and each of these pieces. This was imperative to avoid causing friction and complicating the validation model. A load cell was then purchased and used to measure the pre-load being applied to the testbed. It was also required to collect data for 3 axes at 3 different locations 120° apart on the center-mass. This configuration was required to be placed on the cylindrical face and toward the top and bottom of the center-mass. Slots were milled out in the middle of each sensor block with tight tolerances for mounting each of the accelerometers at 120° apart. This same approach was taken to mount 2 accelerometers on the baseplate at 180° apart. Holes were milled out of the corners of each accelerometer mounting slot to relieve the corner and allow for a flat edge at the back of each slot. Sensor blocks were mounted into the center-mass using 6-32 bolts at the required locations. This initial testbed design was fabricated at the TA53 machine shop.
The objective of the Delivery Environments (DE) Testbeds to Reduce Uncertainties in Simulations and Tests (TRUST) project is to support the efficient and responsive development of experimental, modeling, and simulation capabilities for future systems by developing representative testbeds that can be exercised more easily and efficiently than a WR-like assembly for the purposes uncertainty quantification. The testbeds are intended to be experimentally exercised in current and future relevant engineering environments with complementary modeling and simulation.
The following assessment evaluates the efficacy of the control script for carrying out a linear signal to generate a linear mechanical response of the system: the TRUST nonlinear dynamics (TRUSTND) testbed. The hardware has three main components: the controller (NI PXIe 8861) embedded in the National Instruments chassis (NI PXIe 1092), the signal amplifier (The Modal Shop Linear Power Amplifier 2050E09) to amplify the output of a custom LabVIEW script, and a shaker (The Modal Shop Electrodynamic Exciter 2075E) where wave spring specimens live and are tested within the attached aluminum testbed (center mass). Different kinds of signals (single-tone, swept frequency and white noise) were sent through this hardware in the E-1 lab space at TA-53. First, the linearity of the setup was tested by specifying a sine wave in the control script and checking the quality of the oscillations in a wave spring (McMaster-Carr 9714K19 [1]). When passing a sine wave through the amplifier was succeeded, the frequency response of the wave spring was tested with white noise for comparison to finite-element model predictions. The testbed outputs a repeatable, linear response to the input excitation when used with the wave spring. These measurements reduce the uncertainty associated with the testbed and test procedure to better characterize uncertainty associated with material behavior (SX358 foam samples) in future tests
The objective of the Delivery Environments (DE) Testbeds to Reduce Uncertainties in Simulations and Tests (TRUST) work package is to quantify and help increase confidence in specific areas of computational and experimental capabilities that are applicable to current and future delivery environments. More complete quantification of confidence in experimental and computational capabilities and the sufficient increase of confidence in those capabilities is critical to improving weapons engineering design, qualification, and assessment efforts that are critical to the current and future stockpile. Staff development will include cross-discipline training to provide engineers with experience in both numerical simulations and experimental methods. This work will use and provide feedback on analysis tools and experimental results databases for efficient and responsive engineering which are currently under development: engineering common model framework (ECMF), engineering quantification of margins and uncertainties (EQMU), and the test information management system (TIMS).
The objective of the Delivery Environments (DE) Testbeds to Reduce Uncertainty in Simulations and Tests (TRUST) work package, is to quantify and help increase confidence in specific areas of computational and experimental capabilities that are applicable to current and future delivery environments. More complete quantification of confidence in experimental and computational capabilities and the sufficient increase of confidence in those capabilities is critical to improving weapons engineering design, qualification, and assessment efforts that are relevant to the current and future stockpile. This work will use and provide feedback on analysis tools and experimental results databases for efficient and responsive engineering which are currently under development: engineering common model framework (ECMF), engineering quantification of margins and uncertainties (EQMU), and the test information management system (TIMS). Additionally, this work is being used to test a newly available framework in W-13, Weapons Analysis for Verified Engineering Simulation (WAVES).
There is an increasing aspiration to utilize machine learning (ML) for various tasks of relevance to national security. ML models have thus far been mostly applied to tasks and domains that, while impactful, have sufficient volume of data. For predictive tasks of national security relevance, ML models of great capacity (ability to approximate nonlinear trends in input-output maps) are often needed to capture the complex underlying physics. However, scientific problems of relevance to national security are often accompanied by various sources of sparse and/or incomplete data, including experiments and simulations, across different regimes of operation, of varying degrees of fidelity, and include noise with different characteristics and/or intensity. State-of-the-art ML models, despite exhibiting superior performance on the task and domain they were trained on, may suffer detrimental loss in performance in such sparse data environments. This report summarizes the results of the Laboratory Directed Research and Development project entitled Trust-Enhancing Probabilistic Transfer Learning for Sparse and Noisy Data Environments. The objective of the project was to develop a new transfer learning (TL) framework that aims to adaptively blend the data across different sources in tackling one task of interest, resulting in enhanced trustworthiness of ML models for mission- and safety-critical systems. The proposed framework determines when it is worth applying TL and how much knowledge is to be transferred, despite uncontrollable uncertainties. The framework accomplishes this by leveraging concepts and techniques from the fields of Bayesian inverse modeling and uncertainty quantification, relying on strong mathematical foundations of probability and measure theories to devise new uncertainty-aware TL workflows.
The objective of the Delivery Environments (DE) Testbeds to Reduce Uncertainty in Simulations and Tests (TRUST) project is to quantify and help increase confidence in specific areas of computational and experimental capabilities that are applicable to the development, on-target assessment, and qualification of current and future delivery environments. More complete quantification of confidence in experimental and computational capabilities and the sufficient increase of confidence in those capabilities is critical to improving weapons engineering design, qualification, and assessment efforts that are critical to the current and future stockpile. This work uses and provides feedback on analysis tools and experimental results databases for efficient and responsive engineering which are currently under development.
Building Trust: How to Handle the Challenging Conversations with CCUS Stakeholders
This is one of several manuscripts documenting the outcomes of the Trust in VUI project. We will request sponsor approval, and then plan to submit to the ESARDA Bulletin.
The trusted inertial terrain-aided navigation (TITAN) algorithm leverages an airborne vertical synthetic aperture radar to measure the range to the closest ground points along several prescribed iso-Doppler contours. These TITAN minimum-range, prescribed-Doppler measurements are the result of a constrained nonlinear optimization problem whose optimization function and constraints both depend on the radar position and velocity. Owing to the complexity of this measurement definition, analysis of the TITAN algorithm is lacking in prior work. This publication offers such an analysis, making the following three contributions: (1) an analytical solution to the TITAN constrained optimization measurement problem, (2) a derivation of the TITAN measurement function Jacobian, and (3) a derivation of the Cramér-Rao lower bound on the estimated position and velocity error covariance. These three contributions are verified via Monte Carlo simulations over synthetic terrain, which further reveal two remarkable properties of the TITAN algorithm: (1) the along-track positioning errors tend to be smaller than the cross-track positioning errors, and (2) the cross-track positioning errors are independent of the terrain roughness.
A new method for unconstrained global function optimization, acronymed TRUST, is introduced.
The Global Positioning System (GPS) uses a network of satellites to calculate the position of a receiver over time. This technology has revolutionized a wide range of safety-critical industries and leisure applications. These systems provide diverse benefits; supplementing the users existing navigation skills and reducing the uncertainty that often characterizes many route planning tasks. GPS applications can also help to reduce workload by automating tasks that would otherwise require finite cognitive and perceptual resources. However, the operation of these systems has been identified as a contributory factor in a range of recent accidents. Users often come to rely on GPS applications and, therefore, fail to notice when they develop faults or when errors occur in the other systems that use the data from these systems. Further accidents can stem from the over confidence that arises when users assume automated warnings will be issued when they stray from an intended route. Unless greater attention is paid to the role of trust and interaction in GPS applications then there is a danger that we will see an increasing number of these failures as positioning technologies become integral in the functioning of increasing numbers of applications.
NASA's Earth Observing System Data and Information System (EOSDIS) has been in operation since August 1994, managing most of NASA's Earth science data from satellites, airborne sensors, filed campaigns and other activities. Having been designated by the Federal Government as a project responsible for production, archiving and distribution of these data through its Distributed Active Archive Centers (DAACs), the Earth Science Data and Information System Project (ESDIS) is responsible for EOSDIS, and is legally bound by the Office of Management and Budgets circular A-130, the Federal Records Act. It must follow the regulations of the National Institute of Standards and Technologies (NIST) and National Archive and Records Administration (NARA). It must also follow the NASA Procedural Requirement 7120.5 (NASA Space Flight Program and Project Management). All these ensure that the data centers managed by ESDIS are trustworthy from the point of view of efficient and effective operations as well as preservation of valuable data from NASA's missions. Additional factors contributing to this trust are an extensive set of internal and external reviews throughout the history of EOSDIS starting in the early 1990s. Many of these reviews have involved external groups of scientific and technological experts. Also, independent annual surveys of user satisfaction that measure and publish the American Customer Satisfaction Index (ACSI), where EOSDIS has scored consistently high marks since 2004, provide an additional measure of trustworthiness. In addition, through an effort initiated in 2012 at the request of NASA HQ, the ESDIS Project and 10 of 12 DAACs have been certified by the International Council for Science (ICSU) World Data System (WDS) and are members of the ICSUWDS. This presentation addresses questions such as pros and cons of the certification process, key outcomes and next steps regarding certification. Recently, the ICSUWDS and Data Seal of Approval (DSA) organizations merged their Core Trustworthy Data Repositories Requirements and require that members be recertified every three years. Given the rigor with which NASA manages the ESDIS Project and the DAACs, the recertification through WDSDSA, while involving some additional work, is a relatively simple process.
A poster for the Pathways Student Showcase that includes a description of what Zero Trust Network Security is, the rough timeline for the initial stages of implementation here at KSC, who I am, and what I'm working on.
Deep neural networks are used increasingly for perception and decision-making in UAVs. For example, they can be used to recognize objects from images and decide what actions the vehicle should take. While deep neural networks can perform very well at complex tasks, their decisions may be unintuitive to a human operator. When a human disagrees with a neural network prediction, due to the black box nature of deep neural networks, it can be unclear whether the system knows something the human does not or whether the system is malfunctioning. This uncertainty is problematic when it comes to ensuring safety. As a result, it is important to develop technologies for explaining neural network decisions for trust and safety. This paper explores a modification to the deep neural network classification layer to produce both a predicted label and an explanation to support its prediction. Specifically, at test time, we replace the final output layer of the neural network classifier by a k-nearest neighbor classifier. The nearest neighbor classifier produces 1) a predicted label through voting and 2) the nearest neighbors involved in the prediction, which represent the most similar examples from the training dataset. Because prediction and explanation are derived from the same underlying process, this approach guarantees that the explanations are always relevant to the predictions. We demonstrate the approach on a convolutional neural network for a UAV image classification task. We perform experiments using a forest trail image dataset and show empirically that the hybrid classifier can produce intuitive explanations without loss of predictive performance compared to the original neural network. We also show how the approach can be used to help identify potential issues in the network and training process.