SEARCH · Search NASA

Results for “task tuning”

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 19 records

All sources NASA NTRS DOE OSTI

NASA NTRSDate not supplied

Natural Language Processing Analysis of Notices to Airmen for Air Traffic Management Optimization

With new emerging technologies in the field of NLP, we explore their applications to digitize and analyze heritage Air Traffic Management (ATM) documents for planning and optimizing airspace operations. Specifically, this research focuses on harvesting semi-structured or un-structured information contained in Notices to Airmen (NOTAMs). Using NLP and other advanced data analytics, we will construct a data-driven framework which facilitates finding language patterns and the use of pretrained language models for classification and extraction of useful airspace constraints and restrictions. These may lead to tools that assist airspace users in understanding the constraints more efficiently, contributing to better route planning and safer execution. This paper explores three workflows entailing different NLP tasks. First, unsupervised techniques like word embedding and topic modeling are used for pattern finding and document classification. Second, a dataset is created by extracting information from the semi-structured NOTAM format as metadata for categorizing, visualizing, and extracting key entities driving NOTAM content. Third, modern pre-built deep learning based transformer models such as BERT, RoBERTa, and XLNet are evaluated on the question answering task, an even more robust approach to information extraction, as well as their respective fine-tuning tasks. In this work we include various performance metrics for the trained models to evaluate both accuracy and precision and we show that the models can be generalized for their respective tasks. The research work developed shows promise in uncovering trends in digital NOTAMs in the NAS and also offers a new framework for digitizing and inferring insights from free-form legacy NOTAMs, that are yet to be digitized.

Natural Language Processing↗

NASA NTRSDate not supplied

Natural Language Processing (NLP) Analysis of NOTAMs for Air Traffic Management Optimization

Natural Language Processing↗

NASA NTRS2013

Simulation System Fidelity Assessment at the Vertical Motion Simulator

Fidelity is a word that is often used but rarely understood when talking about groundbased simulation. Assessing the cueing fidelity of a ground based flight simulator requires a comparison to actual flight data either directly or indirectly. Two experiments were conducted at the Vertical Motion Simulator using the GenHel UH-60A Black Hawk helicopter math model that was directly compared to flight data. Prior to the experiment the simulator s motion and visual system frequency responses were measured, the aircraft math model was adjusted to account for the simulator motion system delays, and the motion system gains and washouts were tuned for the individual tasks. The tuned motion system fidelity was then assessed against the modified Sinacori criteria. The first experiments showed similar handling qualities ratings (HQRs) to actual flight for a bob-up and sidestep maneuvers. The second experiment showed equivalent HQRs between flight and simulation for the ADS33 slalom maneuver for the two pilot participants. The ADS33 vertical maneuver HQRs were mixed with one pilot rating the flight and simulation the same while the second pilot rated the simulation worse. In addition to recording HQRs on the second experiment, an experimental Simulation Fidelity Rating (SFR) scale developed by the University of Liverpool was tested for applicability to engineering simulators. A discussion of the SFR scale for use on the Vertical Motion Simulator is included in this paper.

Beard, Steven D.↗

NASA NTRS2003

Software Searches for Better Spacecraft-Navigation Models

ADAPT is a computer program that searches for better mathematical models for spacecraft navigation. The task of tuning trajectory-determination models for interplanetary navigation is complex, requiring an intensive search of multiple dynamical and nondynamical models that yield trajectory solutions with minimal errors. By automating the search, ADAPT eases the task of human analysts and enables them to consider wider ranges of potential solutions. ADAPT uses genetic algorithms to search a range of relevant parameters in a user-selected design space to arrive at values for those parameters that best fit the measured spacecraft-tracking data. The user s guide for ADAPT reviews the theoretical basis of the program and presents two example applications. One example is that of selecting a solar-radiation model for the Mars Pathfinder (MPF) mission using MPF tracking data and an extended Kalman filter from prior spacecraft-navigation software. The second example is of the use of tracking data from the Stardust spacecraft mission combined with a pseudo-epoch-state batch filter and an empirical small-forces model to find improved impulse models for use during Stardust attitude adjustments.

Ely, Todd↗

NASA NTRS2010

In Vitro Disease Model of Microgravity Conditioning on Human Energy Metabolism

NASA and its partners are committed to introducing appropriate new technology to enable learning and living safely beyond the Earth for extended periods of time in a sustainable and possibly indefinite manner. In the responsible acquisition of that goal, life sciences is tasked to tune and advance current medical technology to prepare for human health and wellness in the space environment. The space environment affects the condition and function of biological systems from organ level function to shape of individual organelles. The objective of this paper is to study the effect of microgravity on kinetics of drug metabolism. This fundamental characterization is meaningful to (1) scientific understanding of the response of biology to microgravity and (2) clinical dosing requirements and pharmacological thresholds during long term manned space exploration. Metabolism kinetics of the anti-nausea drug promethazine (PMZ) were determined by an in vitro ground model of 3-dimensional aggregates of human hepatocytes conditioned to weightlessness using a rotating wall bioreactor. The authors observed up-regulated PMZ conversion in model microgravity conditions and attribute this to effect to model microgravity conditioning acting on metabolic mechanisms of the cells. Further work is necessary to determine which particular cellular mechanisms are governing the experimental observations, but the authors conclude kinetics of drug metabolism are responsive to gravitational fields and further study of this sensitivity would improve dosing of pharmaceuticals to persons exposed to a microgravity environment.

Snyder, Jessica↗

NASA NTRSDate not supplied

SafeAeroBERT: Towards a Safety-Informed Aerospace-Specific Language Model

As aviation systems continue to operate with high traffic, large amounts of documents containing safety-relevant data continue to be generated via reporting systems such as the Aviation Safety Reporting System (ASRS). Advanced natural language processing techniques, specifically pre-trained language models, have shown great success in domain-specific applications; however, the text in aviation safety reports is inundated with jargon and thus not fully utilized by general pre-trained models. In this research, we work towards developing a safety-informed aerospace-specific language model by pre-training a Bidirectional Encoder Representations from Transformer (BERT) model on reports from the Aviation Safety Reporting System and the National Transportation Safety Board. The resulting model, called SafeAeroBERT, is fine-tuned for the specific task of document classification, and can be further tuned for named-entity recognition, relation detection, information retrieval, and summarization. Results from the classification task are compared between SafeAeroBERT, the base BERT, and SciBERT models and show SafeAeroBERT outperforms the general BERT and SciBERT on classifying reports about weather and procedure. SafeAeroBERT can be used on custom tasks, not limited to document classification, and is intended to aid an intelligent knowledge manager for safety report repositories.

Aviation↗

NASA NTRS2023

SafeAeroBERT: Towards a Safety-Informed Aerospace-Specific Language Model

As aviation systems continue to operate with high traffic, large amounts of documents containing safety-relevant data continue to be generated via reporting systems such as the ASRS. Advanced natural language processing techniques, specifically pre-trained language models, have shown great success in domain-specific applications; however, the text in aviation safety reports is inundated with jargon and thus not fully utilized by general pre-trained models. In this research, we work towards developing a safety-informed aerospace-specific language model by pre-training a Bidirectional Encoder Representations from Transformer (BERT) model on reports from the Aviation Safety Reporting System and the National Transportation Safety Board. The resulting model, called SafeAeroBERT, is fine-tuned for the specific task of document classification, and can be further tuned for named-entity recognition, relation detection, information retrieval, and summarization. Results from the classification task are compared between SafeAeroBERT, the base BERT, and SciBERT models and show SafeAeroBERT outperforms the general BERT and SciBERT on classifying reports about human factors, aircraft, and procedure. SafeAeroBERT can be used on custom tasks, not limited to document classification, and is intended to aid an intelligent knowledge manager for safety report repositories.

Aviation↗

NASA NTRS2000

Autonomous Performance Monitoring System: Monitoring and Self-Tuning (MAST)

Maintaining the long-term performance of software onboard a spacecraft can be a major factor in the cost of operations. In particular, the task of controlling and maintaining a future mission of distributed spacecraft will undoubtedly pose a great challenge, since the complexity of multiple spacecraft flying in formation grows rapidly as the number of spacecraft in the formation increases. Eventually, new approaches will be required in developing viable control systems that can handle the complexity of the data and that are flexible, reliable and efficient. In this paper we propose a methodology that aims to maintain the accuracy of flight software, while reducing the computational complexity of software tuning tasks. The proposed Monitoring and Self-Tuning (MAST) method consists of two parts: a flight software monitoring algorithm and a tuning algorithm. The dependency on the software being monitored is mostly contained in the monitoring process, while the tuning process is a generic algorithm independent of the detailed knowledge on the software. This architecture will enable MAST to be applicable to different onboard software controlling various dynamics of the spacecraft, such as attitude self-calibration, and formation control. An advantage of MAST over conventional techniques such as filter or batch least square is that the tuning algorithm uses machine learning approach to handle uncertainty in the problem domain, resulting in reducing over all computational complexity. The underlying concept of this technique is a reinforcement learning scheme based on cumulative probability generated by the historical performance of the system. The success of MAST will depend heavily on the reinforcement scheme used in the tuning algorithm, which guarantees the tuning solutions exist.

Peterson, Chariya↗

NASA NTRS1992

SKYMAP: Exploring the Universe in software

SKYMAP is a computer program which produces maps of arbitrary portions of the sky in a variety of projections and coordinate systems. Over the past ten years it has been used to produce finder charts for occultations by planets, display scan and image data from the Spacelab 2 Infrared Telescope, and make maps of fields for astronomical observations at X-ray, optical, infrared, and radio wavelengths. It can display multiple source catalogs, including the HST Guide Star Catalog, as well as solar system objects with astrometric accuracy using the JPL DE-130 ephemeris or tabulated positions. SKYMAP can be tuned to a specific task using an ASCII parameter file which controls how information is displayed on any Tektronix-compatible graphics display. The program contains a variety of interactive graphic and image processing features and has been ported to a variety of computer systems. A recent project visually demonstrates source density variation in various commonly-used all-sky catalogs.

Mink, D. J.↗

NASA NTRS2004

Addressing Control Research Issues Leading to Piloted Simulations in Support of the IFCS F-15

This report summarizes the research effort by a team of researchers at West Virginia University in support of the NASA Intelligent Flight Control System (IFCS) F-15 program. In particular, WVU researchers assisted NASA Dryden researchers in the following technical tasks leading to piloted simulation of the 'Gen_2' IFCS control laws. Task #1- Performance comparison of different neural network (NN) augmentation for the Dynamic Inversion (DI) -based VCAS 'Gen_2' control laws. Task #2- Development of safety monitor criteria for transition to research control laws with and without failure during flight test. Task #3- Fine-tuning of the 'Gen_2' control laws for cross-coupling reduction at post-failure conditions. Matlab/Simulink-based simulation codes were provided to the technical monitor on a regular basis throughout the duration of the project. Additional deliverables for the project were Power Point-based slides prepared for different project meetings. This document provides a description of the methodology and discusses the general conclusions from the simulation results.

Napolitano, Marcello↗

NASA NTRS2005

Software Would Largely Automate Design of Kalman Filter

Embedded Navigation Filter Automatic Designer (ENFAD) is a computer program being developed to automate the most difficult tasks in designing embedded software to implement a Kalman filter in a navigation system. The most difficult tasks are selection of error states of the filter and tuning of filter parameters, which are timeconsuming trial-and-error tasks that require expertise and rarely yield optimum results. An optimum selection of error states and filter parameters depends on navigation-sensor and vehicle characteristics, and on filter processing time. ENFAD would include a simulation module that would incorporate all possible error states with respect to a given set of vehicle and sensor characteristics. The first of two iterative optimization loops would vary the selection of error states until the best filter performance was achieved in Monte Carlo simulations. For a fixed selection of error states, the second loop would vary the filter parameter values until an optimal performance value was obtained. Design constraints would be satisfied in the optimization loops. Users would supply vehicle and sensor test data that would be used to refine digital models in ENFAD. Filter processing time and filter accuracy would be computed by ENFAD.

Chuang, Jason C. H.↗

NASA NTRS2009

Updating the Finite Element Model of the Aerostructures Test Wing using Ground Vibration Test Data

Improved and/or accelerated decision making is a crucial step during flutter certification processes. Unfortunately, most finite element structural dynamics models have uncertainties associated with model validity. Tuning the finite element model using measured data to minimize the model uncertainties is a challenging task in the area of structural dynamics. The model tuning process requires not only satisfactory correlations between analytical and experimental results, but also the retention of the mass and stiffness properties of the structures. Minimizing the difference between analytical and experimental results is a type of optimization problem. By utilizing the multidisciplinary design, analysis, and optimization (MDAO) tool in order to optimize the objective function and constraints; the mass properties, the natural frequencies, and the mode shapes can be matched to the target data to retain the mass matrix orthogonality. This approach has been applied to minimize the model uncertainties for the structural dynamics model of the Aerostructures Test Wing (ATW), which was designed and tested at the National Aeronautics and Space Administration (NASA) Dryden Flight Research Center (DFRC) (Edwards, California). This study has shown that natural frequencies and corresponding mode shapes from the updated finite element model have excellent agreement with corresponding measured data.

Lung, Shun-fat↗

NASA NTRS2009

Updating the Finite Element Model of the Aerostructures Test Wing Using Ground Vibration Test Data

Improved and/or accelerated decision making is a crucial step during flutter certification processes. Unfortunately, most finite element structural dynamics models have uncertainties associated with model validity. Tuning the finite element model using measured data to minimize the model uncertainties is a challenging task in the area of structural dynamics. The model tuning process requires not only satisfactory correlations between analytical and experimental results, but also the retention of the mass and stiffness properties of the structures. Minimizing the difference between analytical and experimental results is a type of optimization problem. By utilizing the multidisciplinary design, analysis, and optimization (MDAO) tool in order to optimize the objective function and constraints; the mass properties, the natural frequencies, and the mode shapes can be matched to the target data to retain the mass matrix orthogonality. This approach has been applied to minimize the model uncertainties for the structural dynamics model of the aerostructures test wing (ATW), which was designed and tested at the National Aeronautics and Space Administration Dryden Flight Research Center (Edwards, California). This study has shown that natural frequencies and corresponding mode shapes from the updated finite element model have excellent agreement with corresponding measured data.

Lung, Shun-Fat↗

NASA NTRS2020

Shedding Light on Microbial Dark Matter with A Universal Language of Life

The majority of microbial genomes have yet to be cultured, and most proteins predicted from microbial genomes or sequenced from the environment cannot be functionally annotated. As a result, current computational approaches to describe microbial systems rely on incomplete reference databases that cannot adequately capture the full functional diversity of the microbial tree of life, limiting our ability to model high-level features of biological sequences. The scientific community needs a means to capture the functionally and evolutionarily relevant features underlying biology, independent of our incomplete reference databases. Such a model can form the basis for transfer learning tasks, enabling downstream applications in environmental microbiology, medicine, and bioengineering. Here we present LookingGlass, a deep learning model capturing a “universal language of life”. LookingGlass encodes contextually-aware, functionally and evolutionarily relevant representations of short DNA reads, distinguishing reads of disparate function, homology, and environmental origin. We demonstrate the ability of LookingGlass to be fine-tuned to perform a range of diverse tasks: to identify novel oxidoreductases, to predict enzyme optimal temperature, and to recognize the reading frames of DNA sequence fragments. LookingGlass is the first contextually-aware, general purpose pre-trained “biological language” representation model for short-read DNA sequences. LookingGlass enables functionally relevant representations of otherwise unknown and unannotated sequences, shedding light on the microbial dark matter that dominates life on Earth.

A Hoarfrost↗

NASA NTRSDate not supplied

Towards an Aviation Large Language Model by Fine-tuning and Evaluating Transformers

In the aviation domain, there are many applications for machine learning and artificial intelligence tools that utilize natural language. For example, there is a desire to know the commonalities in written safety reports such as voluntary post incidents reports or aerial wildfire operations reports to better understand the risks present. Another use-case is the possibility of extracting airspace procedures and constraints currently written in documents such as Letters of Agreement. These applications can benefit from the use of state-of-the-art natural language processing techniques when adapted to the language/phraseology specific to the aviation domain. This paper evaluates the viability of adaptation of NLP tools to the aviation domain by fine-tuning transformer based models using aviation data sets. In 2018, a novel language model based on neural units (also called transformers) was created and became known as “Bidirectional Encoder Representations from Transformers” or BERT. This architecture combined with large amounts of English training data and innovative semi-supervised training tasks set the standard for what would later emerge as Large Language Models. The performance of these models was further improved by hyperparameter tuning and refinement of the semi-supervised training task and resulted in “Robustly Optimized BERT Pre-training Approach through hyperparameter tuning” or RoBERTa models. These pre-trained Large Language Models proved to be useful for a wide variety of natural language processing tasks such as text classification and question answering through a process called fine-tuning. The transformer architecture with pre-trained weights served as the basis with the last few layers replaced with layers fine-tuned to perform a new task e.g., a layer that provides a label for the entire input text. This process of fine-tuning can also be used to adapt the models to new domains; e.g., BioBERT started with the pre-trained BERT model and was completed by additional fine-tuning and training on biomedical documents. Transformer-based architectures can also be used to create rich representations of text called embeddings which can serve as the input to other machine learning models. This allows simpler algorithms such as logistic regression to use context-rich representations of the text while still remaining quick to train and evaluate. In the world of aviation, there is a growing demand for natural language processing and understanding but the domain presents unique challenges. Due to the technical content (and specialized language) of most aviation documents, fine-tuning pre-trained Large Language Models to specific tasks has not met the benchmark on natural language processing tasks set by simpler models trained from scratch on the data. To address this deficiency, this paper evaluates the improvements from fine-tuning a Large Language Model on a large set of aviation documents using the original semi-supervised training tasks before performing specific natural language tasks. In fine-tuning, a domain-specific dataset is used on the original training task but with the pre-trained Large Language Model instead of starting from a random initialization. This approach allows the model to be adapted to the specific domain language without discarding the information gained from training on general English data. This paper utilized two major dataset types to train and assess the RoBERTa fine-tuning performance. The first are 7,057 Letters of Agreement which are Federal Aviation Administration (FAA) documents that formalize airspace operations across the national airspace system. They contain many examples of ‘aviation English’ using domain specific terminology and phrasing which serves as a representative basis to perform the semi-supervised fine-tuning. The second type is the 494 document classification labels to be used for evaluation. This down-stream evaluation aims to show the performance of the fine-tuned model, better understand how much data is needed for an effective fine-tuning, and how fine-tuning can be adapted for different applications in-the domain. After semi-supervised training, evaluation begins by encoding the documents for classification using the fine-tuned RoBERTa model. Then a logistic regression classifier is trained to label the document type and compared against our ground truth labels. This currently leads to a 82.8% accuracy on 10-fold cross validation showing improvement over baseline RoBERTa which achieved 81.0%. We plan to measure the improvements on additional tasks and it is expected that these improvements will lead to more robust models that can tackle the natural language processing challenges present in aviation datasets.

ATM↗

NASA NTRS2014

Range-Specific High-Resolution Mesoscale Model Setup: Data Assimilation

Mesoscale weather conditions can have an adverse effect on space launch, landing, and ground processing at the Eastern Range (ER) in Florida and Wallops Flight Facility (WFF) in Virginia. During summer, land-sea interactions across Kennedy Space Center (KSC) and Cape Canaveral Air Force Station (CCAFS) lead to sea breeze front formation, which can spawn deep convection that can hinder operations and endanger personnel and resources. Many other weak locally driven low-level boundaries and their interactions with the sea breeze front and each other can also initiate deep convection in the KSC/CCAFS area. Some of these other boundaries include the Indian River breeze front, Banana River breeze front, outflows from previous convection, horizontal convective rolls, convergence lines from other inland bodies of water such as Lake Okeechobee, the trailing convergence line from convergence of sea breeze fronts due to the shape of Cape Canaveral, frictional convergence lines from the islands in the Bahamas, convergence lines from soil moisture differences, convergence lines from cloud shading, and others. All these subtle weak boundary interactions often make forecasting of operationally important weather very difficult at KSC/CCAFS during the convective season (May-Oct). These convective processes often build quickly, last a short time (60 minutes or less), and occur over small distances, all of which also poses a significant challenge to the local forecasters who are responsible for issuing weather advisories, watches, and warnings. Surface winds during the transition seasons of spring and fall pose the most difficulties for the forecasters at WFF. They also encounter problems forecasting convective activity and temperature during those seasons. Therefore, accurate mesoscale model forecasts are needed to aid in their decision making. Both the ER and WFF would benefit greatly from high-resolution mesoscale model output to better forecast a variety of unique weather phenomena. Global and national scale models cannot properly resolve important local-scale weather features at each location due to their horizontal resolutions being much too coarse. Therefore, a properly tuned model at a high resolution is needed to provide improved capability. This task is a multi-year effort in which the Applied Meteorology Unit (AMU) will tune the Weather Research and Forecasting (WRF) model individually for each range. The goal of the first year, the results of which are in this report, was to tune the WRF model based on the best model resolution and run time while using reasonable computing capabilities. To accomplish this, the ER and WFF supported the tasking of the AMU to perform a number of sensitivity tests in order to determine the best model configuration for operational use at each of the ranges to best predict winds, precipitation, and temperature (Watson 2013). This task is a continuation of that work and will provide a recommended local data assimilation (DA) and numerical forecast model design optimized for the ER and WFF to support space launch activities. The model will be optimized for local weather challenges at both ranges.

Global and national scale models↗

NASA NTRS1981

Modal analysis using a Fourier analyzer, curve-fitting, and modal tuning

The proposed modal test program differs from single-input methods in that preliminary data may be acquired using multiple inputs, and modal tuning procedures may be employed to define closely spaced frquency modes more accurately or to make use of frequency response functions (FRF's) which are based on several input locations. In some respects the proposed modal test proram resembles earlier sine-sweep and sine-dwell testing in that broadband FRF's are acquired using several input locations, and tuning is employed to refine the modal parameter estimates. The major tasks performed in the proposed modal test program are outlined. Data acquisition and FFT processing, curve fitting, and modal tuning phases are described and examples are given to illustrate and evaluate them.

Craig, R. R., Jr.↗

NASA NTRS1998

Direct Numerical Simulations of Boundary Layer Transition on a Flat Plate

In recent years the techniques of computational fluid dynamics (CFD) have been used to compute flows associated with geometrically complex configurations. However, success in terms of accuracy and reliability has been limited to cases where the effects of turbulence and transition could be modeled in a straightforward manner. Even in simple flows, the accurate computation of skin friction and heat transfer using existing turbulence models has proved to be a difficult task, one that has required extensive fine-tuning of the turbulence models used. In more complex flows (for example, in turbomachinery flows in which vortices and wakes impinge on airfoil surfaces causing periodic transitions from laminar to turbulent flow) the development of a model that accounts for all scales of turbulence and predicts the onset of transition is an extremely difficult task. Fortunately, current trends in computing suggest that it may be possible to perform direct simulations of turbulence and transition at moderate Reynolds numbers in some complex cases in the near future. This presentation will focus on direct simulations of transition and turbulence using high-order accurate finite-difference methods. The advantage of the finite-difference approach over spectral methods is that complex geometries can be treated in a straightforward manner. Additionally, finite-difference techniques are the prevailing methods in existing application codes. An application of high-order-accurate finite-difference methods to direct simulations of transition and turbulence in a spatially evolving boundary layer subjected to high levels of freestream turbulence will be presented.

Rai, Man Mohan↗