SEARCH · Search NASA
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.
This paper presents a method for extracting information from SAR images using the wavelet transform. It is shown experimentally that the shape of the wavelet can be chosen in order to extract different features in the image.
Traditional protection schemes face significant challenges when applied to microgrids with high penetrations of renewables with inverter-based resources (IBRs). The proliferation of advanced sensing and communication technologies has generated copious data, offering an opportunity to overcome these limitations using data-driven machine learning approaches. This work proposes a novel approach based on a support vector machine (SVM) for detecting faults within a 100% renewable microgrid. The approach encompasses a systematic offline training stage for the development of a linear SVM-based fault detection algorithm. This process covers offline data collection from the microgrid under study, the extraction of features such as positive- and negative-sequence components and the total harmonic distortion of the voltage and current measurements of the relays, and the design of the linear SVM-based classifier. During the online implementation, however, different classifiers can exhibit asynchronicity in detecting the fault inception at different subcycle-to-cycle period-level delays. To circumvent this asynchronicity issue, a separate algorithm is developed for each relay to estimate the fault inception time as close to the real fault time. The performance of the proposed SVM-based synchronized fault detection method is evaluated using online time-domain simulation studies on a microgrid test system. The results corroborate the reliability of the fault detection scheme when tested under various fault cases (fault types, locations, and impedances) and non-fault cases during both grid-tied and islanded operation modes.
This paper presents an algorithm to find the correspondences of points representing dominant feature in robot stereo vision. The algorithm consists of two main steps: dominant point extraction and dominant point matching. In the feature extraction phase, the algorithm utilizes the widely used Moravec Interest Operator and two other operators: the Prewitt Operator and a new operator called Gradient Angle Variance Operator. The Interest Operator in the Moravec algorithm was used to exclude featureless areas and simple edges which are oriented in the vertical, horizontal, and two diagonals. It was incorrectly detecting points on edges which are not on the four main directions (vertical, horizontal, and two diagonals). The new algorithm uses the Prewitt operator to exclude featureless areas, so that the Interest Operator is applied only on the edges to exclude simple edges and to leave interesting points. This modification speeds-up the extraction process by approximately 5 times. The Gradient Angle Variance (GAV), an operator which calculates the variance of the gradient angle in a window around the point under concern, is then applied on the interesting points to exclude the redundant ones and leave the actual dominant ones. The matching phase is performed after the extraction of the dominant points in both stereo images. The matching starts with dominant points in the left image and does a local search, looking for corresponding dominant points in the right image. The search is geometrically constrained the epipolar line of the parallel-axes stereo geometry and the maximum disparity of the application environment. If one dominant point in the right image lies in the search areas, then it is the corresponding point of the reference dominant point in the left image. A parameter provided by the GAV is thresholded and used as a rough similarity measure to select the corresponding dominant point if there is more than one point the search area. The correlation is used as a final decision tool when there is still more than one point in the search area. If there is no dominant point in the search area of if the points in the search area are below a correlation threshold, then the dominant point in the reference image is occluded and can not be corresponded. The algorithm has been modeled, implemented and shown to be fast, robust and parallel. The parallelism is created from three main features: locality of the operators; a memory optimization scheme; and the ability to fully parallelize the extraction phase which is the most computational intensive task in the algorithm. The last feature is achieved by performing the extraction phase on the two images simultaneously.
We present the development of advanced automatic target recognition (ATR) algorithms for the hair follicles identification in digital skin images to accurately direct the laser beam to remove the hair. The ATR system first performs a wavelet filtering to enhance the contrast of the hair features in the image. The system then extracts the unique features of the targets and sends the features to an Adaboost based classifier for training and recognition operations. The ATR system automatically classifies the hair, moles, or other skin lesion and provides the accurate coordinates of the intended hair follicle locations. The coordinates can be used to guide a scanning laser to focus energy only on the hair follicles. The intended benefit would be to protect the skin from unwanted laser exposure and to provide more effective skin therapy.
A machine learning inversion method is developed for analyzing scattering functions of mechanically driven polymers and extracting the corresponding feature parameters, which include energy parameters and conformation variables. The polymer is modeled as a chain of fixed-length bonds constrained by bending energy, and it is subject to external forces such as stretching and shear. We generate a data set consisting of random combinations of energy parameters, including bending modulus, stretching and shear force, along with Monte Carlo-calculated scattering functions and conformation variables such as end-to-end distance, radius of gyration and off-diagonal component of the gyration tensor. The effects of the energy parameters on the polymer are captured by the scattering function, and principal component analysis ensures the feasibility of the machine learning inversion. Finally, we train a Gaussian process regressor using part of the data set as a training set and validate the trained regressor for inversion using the rest of the data. The regressor successfully extracts the feature parameters.
This report presents machine learning (ML) analysis of temperature-strain relationships for structural health monitoring of nuclear reactor stainless steel (SS) pipes with the strain gauge sensor directly printed on the pipe with a 3D conformal aerosol jet printer. We investigate correlations for two sensor pairs installed on the same SS304 pipe: commercial K-type thermocouple with a printed gold strain gauge (TC3-SG3), and commercial K-type thermocouple with commercial Kyowa strain gauge (TC0-SG0). The temperature ranges for the sensor pairs TC0-SG0 and TC3-SG3 are 20.00°C to 266.37°C and 39.95°C to 219.28°C respectively. ML algorithms in this study include Linear Regression (baseline method), Ridge Regression, Lasso Regression, and Gradient Boosting. Performance evaluation metrics include Root Mean Square Error (RMSE), Mean Square Error (MSE), Mean Absolute Error (MAE), R 2 Score, and Explained Variance. Using advanced feature engineering techniques, we extracted 27 temperature-based features and 30 strategic inclusion features. The best performance was obtained with the Gradient Boosting method, which achieves prediction accuracy of R 2 = 0.9999 and RMSE = 7.69 μStrain for TC0-SG0, and R 2 = 0.9998 and RMSE = 18.03 μStrain for TC3-SG3. While the temperature-strain correlations are weaker for the gauge directly printed on the pipe than for the commercial strain gauge, deployment-ready performance exceeding industry standards is achieved for both sensor pairs.
Applying standard software engineering practices to neural networks is challenging due to the lack of high-level abstractions describing a neural network’s behavior. To address this challenge, we propose to extract high-level task-specific features from the neural network internal representation, based on monitoring the neural network activations.The extracted feature representations can serve as a link to high-level requirements and can be leveraged to enable fundamental software engineering activities, such as automated testing, debugging, requirements analysis, and formal verification, leading to better engineering of neural networks. Using two case studies, we present initial empirical evidence demonstrating the feasibility of our ideas.
Object localization and its application in tele-autonomous systems are studied. Two object localization algorithms are presented together with the methods of extracting several important types of object features. The first algorithm is based on line-segment to line-segment matching. Line range sensors are used to extract line-segment features from an object. The extracted features are matched to corresponding model features to compute the location of the object. The inputs of the second algorithm are not limited only to the line features. Featured points (point to point matching) and featured unit direction vectors (vector to vector matching) can also be used as the inputs of the algorithm, and there is no upper limit on the number of the features inputed. The algorithm will allow the use of redundant features to find a better solution. The algorithm uses dual number quaternions to represent the position and orientation of an object and uses the least squares optimization method to find an optimal solution for the object's location. The advantage of using this representation is that the method solves for the location estimation by minimizing a single cost function associated with the sum of the orientation and position errors and thus has a better performance on the estimation, both in accuracy and speed, than that of other similar algorithms. The difficulties when the operator is controlling a remote robot to perform manipulation tasks are also discussed. The main problems facing the operator are time delays on the signal transmission and the uncertainties of the remote environment. How object localization techniques can be used together with other techniques such as predictor display and time desynchronization to help to overcome these difficulties are then discussed.
Feature selection has been one of the most important topics in pattern recognition. Although many authors have studied feature selection for parametric classifiers, few algorithms are available for feature selection for nonparametric classifiers. In this paper we propose a new feature selection algorithm based on decision boundaries for nonparametric classifiers. We first note that feature selection for pattern recognition is equivalent to retaining 'discriminantly informative features', and a discriminantly informative feature is related to the decision boundary. A procedure to extract discriminantly informative features based on a decision boundary for nonparametric classification is proposed. Experiments show that the proposed algorithm finds effective features for the nonparametric classifier with Parzen density estimation.
NASA Data Active Archive Centers, or DAACs, ingest, store and distribute data acquired from satellites, ground systems as well as modelling data. These data are organized by the datasets, each presenting collection of files usually associated with the certain mission, instrument, processing level, parameter(s), algorithm and/or model. The number of datasets offered by a single DAAC to the public varies. GES DISC, for example, currently offers for public use approximately ~1,300 datasets. While each publicly offered dataset comes with supporting documentation, it is challenging for novice and even experienced scientists to navigate among the datasets that offer similar parameters to find the datasets for their particular research application. Supplying dataset documentation with the scientific paper citations that refer to that dataset provides means for the dataset users to educate themselves with the application research that dataset is being used in. Collecting citations of the papers that use the datasets for their research yield valuable insights into application areas of those datasets, information about usage of the dataset groups for specific applications and those application topics. It also gives insights into the “deep metrics” of the dataset usage, as opposed to the common metrics of the dataset usage such as number of users who downloaded the dataset files and volumes of downloaded data. Association of a certain scientific paper with the dataset(s) presents a challenge because most of the paper authors do not properly cite the datasets, datasets usually have cryptic names and Digital Object Identifiers (DOIs) that are used for dataset identification were assigned to the datasets only few years ago. Simple Google or online library search do not provide even meaningful fraction of the results when performed by the dataset name or DOI, however they provide too many results when the search is done by more broader terms such as mission and instrument names. Attempts to create an AI system capable to identify dataset in the scientific papers have already been made using neural networks classifiers on the basis of the dataset mission, instrument and variable name. This method was applied to NASA SEDAC, which has 41 datasets in total. In GES DISC there can be as many as ~100 datasets per mission/instrument with some of the datasets consisting of multiple variables so there is a need for more differentiating parameters for dataset identification in the paper. The approach we are currently investigating is creating AI classifiers that are based on multiple dataset features, or keywords, extracted from the NASA Earthdata Common Dataset Repository (CMR). The features are weighted based on how precisely they can identify a dataset. The classifier uses preprocessed paper text as input and searches for the CMR datasets whose feature sets are the closest to the feature sets contained in the paper. The challenges of dataset identification include variety of ways the paper authors describe the datasets in their papers and incomplete tagging of the CMR dataset description (DIFs).
A hierarchical, decentralized diagnostic system is proposed for the Real-Time Diagnostic System component of the Intelligent Control System (ICS) for reusable rocket engines. The proposed diagnostic system has three layers of information processing: condition monitoring, fault mode detection, and expert system diagnostics. The condition monitoring layer is the first level of signal processing. Here, important features of the sensor data are extracted. These processed data are then used by the higher level fault mode detection layer to do preliminary diagnosis on potential faults at the component level. Because of the closely coupled nature of the rocket engine propulsion system components, it is expected that a given engine condition may trigger more than one fault mode detector. Expert knowledge is needed to resolve the conflicting reports from the various failure mode detectors. This is the function of the diagnostic expert layer. Here, the heuristic nature of this decision process makes it desirable to use an expert system approach. Implementation of the real-time diagnostic system described above requires a wide spectrum of information processing capability. Generally, in the condition monitoring layer, fast data processing is often needed for feature extraction and signal conditioning. This is usually followed by some detection logic to determine the selected faults on the component level. Three different techniques are used to attack different fault detection problems in the NASA LeRC ICS testbed simulation. The first technique employed is the neural network application for real-time sensor validation which includes failure detection, isolation, and accommodation. The second approach demonstrated is the model-based fault diagnosis system using on-line parameter identification. Besides these model based diagnostic schemes, there are still many failure modes which need to be diagnosed by the heuristic expert knowledge. The heuristic expert knowledge is implemented using a real-time expert system tool called G2 by Gensym Corp. Finally, the distributed diagnostic system requires another level of intelligence to oversee the fault mode reports generated by component fault detectors. The decision making at this level can best be done using a rule-based expert system. This level of expert knowledge is also implemented using G2.
Model-based diagnosis typically uses analytical redundancy to compare predictions from a model against observations from the system being diagnosed. However this approach does not work very well when it is not feasible to create analytic relations describing all the observed data, e.g., for vibration data which is usually sampled at very high rates and requires very detailed finite element models to describe its behavior. In such cases, features (in time and frequency domains) that contain diagnostic information are extracted from the data. Since this is a computationally intensive process, it is not efficient to extract all the features all the time. In this paper we present an approach that combines the analytic model-based and feature-driven diagnosis approaches. The analytic approach is used to reduce the set of possible faults and then features are chosen to best distinguish among the remaining faults. We describe an implementation of this approach on the Flyable Electro-mechanical Actuator (FLEA) test bed.
The author has identified the following significant results. Preliminary digital analysis of ERTS-1 MSS imagery reveals that the textural features of the imagery are very useful for land use classification. A procedure for extracting the textural features of ERTS-1 imagery is presented and the results of a land use classification scheme based on the textural features are also presented. The land use classification algorithm using textural features was tested on a 5100 square mile area covered by part of an ERTS-1 MSS band 5 image over the California coastline. The image covering this area was blocked into 648 subimages of size 8.9 square miles each. Based on a color composite of the image set, a total of 7 land use categories were identified. These land use categories are: coastal forest, woodlands, annual grasslands, urban areas, large irrigated fields, small irrigated fields, and water. The automatic classifier was trained to identify the land use categories using only the textural characteristics of the subimages; 75 percent of the subimages were assigned correct identifications. Since texture and spectral features provide completely different kinds of information, a significant increase in identification accuracy will take place when both features are used together.
A need exists for digitized information pertaining to linear features such as roads, streams, water bodies and agricultural field boundaries as component parts of a data base. For many areas where this data may not yet exist or is in need of updating, these features may be extracted from remotely sensed digital data. This paper examines two approaches for identifying linear features, one utilizing raw data and the other classified data. Each approach uses a series of data enhancement procedures including derivation of standard deviation values, principal component analysis and filtering procedures using a high-pass window matrix. Just as certain bands better classify different land covers, so too do these bands exhibit high spectral contrast by which boundaries between land covers can be delineated. A few applications for this kind of data are briefly discussed, including its potential in a Universal Soil Loss Equation Model.
We propose that Kuiper Belt samples (in this case comet coma grains from the Jupiter family comet Wild 2) are recognizably different from the bulk of materials in outer belt asteroids, because of their different formation positions and times in the early solar system. We believe this despite similarities found between some Wild 2 grains and components of carbonaceous chondrites (i.e. some CAI and chondrules). Kuiper Belt samples must preserve measurable mineralogical and compositional evidence of formation at unique positions and times in the early solar nebula, and these formational differences must have imparted recognizable special characteristics. We hypothesize that these characteristics include: (1) Unique major element compositional ranges of common astromaterial minerals, especially olivine and pyroxene; (2) Unique minor element compositions of major silicate phases, especially olivine and low-Ca pyroxene; (3) Degree and effects of radiation processing -- including amorphous rims, metal coatings, and Glass with Embedded Metal and Sulfides (GEMS); (4) Presence of abundant presolar silicate grains as recognized by anomalous oxygen in silicates; (5) Oxidation state of the mineral assemblage. We are working our way through all available Wild 2 samples, selecting 1-2 non-consecutive viable TEM grids from each possible extracted Wild 2 grain. We especially prefer TEM grids from grains for which complete mineralogical details have not been published (which is to say the majority of the extracted grains). We are performing a basic mineralogic survey by E-beam techniques, to establish the essential features of the extracted Wild 2 grains. We are making a particular effort to carefully and accurately measure minor elements of olivine and pyroxene, as these minerals are widespread in astromaterials, and comparisons of their compositions will serve to place the Wild 2 silicates in contact with asteroids, meteorites and chondritic interplanetary dust particles processing. We are also making a special effort to search for mineralogical products of aqueous alteration, since their presence would reveal that Wild 2 was once internally heated, a result with dramatic implications for models of early solar system primitive bodies. Thus far carbonates are the only potential evidence for aqueous alteration for Wild 2.
A software system has been developed for prioritizing newly acquired geological data onboard a planetary rover. The system has been designed to enable efficient use of limited communication resources by transmitting the data likely to have the most scientific value. This software operates onboard a rover by analyzing collected data, identifying potential scientific targets, and then using that information to prioritize data for transmission to Earth. Currently, the system is focused on the analysis of acquired images, although the general techniques are applicable to a wide range of data modalities. Image prioritization is performed using two main steps. In the first step, the software detects features of interest from each image. In its current application, the system is focused on visual properties of rocks. Thus, rocks are located in each image and rock properties, such as shape, texture, and albedo, are extracted from the identified rocks. In the second step, the features extracted from a group of images are used to prioritize the images using three different methods: (1) identification of key target signature (finding specific rock features the scientist has identified as important), (2) novelty detection (finding rocks we haven t seen before), and (3) representative rock sampling (finding the most average sample of each rock type). These methods use techniques such as K-means unsupervised clustering and a discrimination-based kernel classifier to rank images based on their interest level.
Background: Future missions to ocean worlds, such as Enceladus and Europa, will attempt to characterize the subsurface seawater chemistry and assess the potential for life. Such missions will be equipped with capabilities to precisely measure volatile isotopes in plumes, atmospheres, and exospheres. Motivation: While large isotopic fractionations can indicate a biological source, there are signatures resulting from abiotic geochemical processes that mimic isotopic biosignatures. While machine learning (ML) has the potential to disentangle competing effects and biotic mimicry, high-dimensional isotope ratio mass spectrometry (IRMS) data is likely to contain noise/irrelevant features and involve complex statistical interactions that make human inference and interpretation difficult. Further, ML predictions with as far-reaching implications as an extraterrestrial biosignature on an ocean world requires the use of interpretable models (i.e., not “black box” models) with physically and mathematically meaningful feature spaces along with false positive diagnostics. Methods: We use volatile CO2 IRMS data of analogue ocean world seawaters to validate an ML approach to provide biogeochemical context for biosignature detection. We employ a feature selection method called nearest-neighbor projected distance regression (NPDR) that detects statistical interactions and helps elucidate the mechanisms of the Random Forest classification models. Results: We train and validate predictive ML models on volatile CO2 IRMS data of analogue ocean world seawaters to predict major salt components (e.g., MgSO4, NaHCO3), pH, ionic strength, and the presence of biosignatures. Features derived from IRMS measurements are augmented with extracted time-series features. Our results show high test accuracy and interpretability, which is increased by interaction network visualization, sample-wise variable importance scores, and single-sample class probability estimates. We demonstrate an ML mission software solution that triggers autonomous data transmission and biogeochemical sample prediction.
A spectral feature design system for high-dimensional multispectral data is described. This system utilizes hard-limited or infinitely clipped optimal transforms and canonical analysis to extract the spectral features for data volume reduction and classification purposes. The design procedure is intended to be application-specific in order to make it maximally effective for each use. Although it could be used in a variety of circumstances, the procedure was designed with satellite data collection in mind, such as will be needed with the High-Resolution Imaging Spectrometer (HIRIS), a land-oriented Earth observational sensor intended for launch in the mid-1990s. The computation required for the steps prior to the actual satellite data collection are straightforward and could be done with readily available subroutines. The procedure is also defined in such a way as to require only very simple onboard calculations. The tests reported provide substantial data volume reduction in the satellite-to-Earth downlink and subsequent computation phases, while maintaining satisfactory classification accuracy.