SEARCH · Search NASA

Results for “Predictors”

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 469 records · Page 26

All sources NASA NTRS DOE OSTI

DOE OSTI2021

MULTICOM2 open-source protein structure prediction system powered by deep learning and distance prediction

Protein structure prediction is an important problem in bioinformatics and has been studied for decades. However, there are still few open-source comprehensive protein structure prediction packages publicly available in the field. In this paper, we present our latest open-source protein tertiary structure prediction system—MULTICOM2, an integration of template-based modeling (TBM) and template-free modeling (FM) methods. The template-based modeling uses sequence alignment tools with deep multiple sequence alignments to search for structural templates, which are much faster and more accurate than MULTICOM1. The template-free (ab initio or de novo) modeling uses the inter-residue distances predicted by DeepDist to reconstruct tertiary structure models without using any known structure as template. In the blind CASP14 experiment, the average TM-score of the models predicted by our server predictor based on the MULTICOM2 system is 0.720 for 58 TBM (regular) domains and 0.514 for 38 FM and FM/TBM (hard) domains, indicating that MULTICOM2 is capable of predicting good tertiary structures across the board. It can predict the correct fold for 76 CASP14 domains (95% regular domains and 55% hard domains) if only one prediction is made for a domain. The success rate is increased to 3% for both regular and hard domains if five predictions are made per domain. Moreover, the prediction accuracy of the pure template-free structure modeling method on both TBM and FM targets is very close to the combination of template-based and template-free modeling methods. This demonstrates that the distance-based template-free modeling method powered by deep learning can largely replace the traditional template-based modeling method even on TBM targets that TBM methods used to dominate and therefore provides a uniform structure modeling approach to any protein. Finally, on the 38 CASP14 FM and FM/TBM hard domains, MULTICOM2 server predictors (MULTICOM-HYBRID, MULTICOM-DEEP, MULTICOM-DIST) were ranked among the top 20 automated server predictors in the CASP14 experiment. After combining multiple predictors from the same research group as one entry, MULTICOM-HYBRID was ranked no. 5. The source code of MULTICOM2 is freely available at https://github.com/multicom-toolbox/multicom/tree/multicom_v2.0 .

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2011

Correlation and prediction uncertainties in the CyberKnife Synchrony respiratory tracking system

Purpose: The CyberKnife uses an online prediction model to improve radiation delivery when treating lung tumors. This study evaluates the prediction model used by the CyberKnife radiation therapy system in terms of treatment margins about the gross tumor volume (GTV). Methods: From the data log files produced by the CyberKnife synchrony model, the uncertainty in radiation delivery can be calculated. Modeler points indicate the tracked position of the tumor and Predictor points predict the position about 115 ms in the future. The discrepancy between Predictor points and their corresponding Modeler points was analyzed for 100 treatment model data sets from 23 de-identified lung patients. The treatment margins were determined in each anatomic direction to cover an arbitrary volume of the GTV, derived from the Modeler points, when the radiation is targeted at the Predictor points. Each treatment model had about 30 min of motion data, of which about 10 min constituted treatment time; only these 10 min were used in the analysis. The frequencies of margin sizes were analyzed and truncated Gaussian normal functions were fit to each direction's distribution. The standard deviation of each Gaussian distribution was then used to describe the necessary margin expansions in each signed dimension in order to achieve the desired coverage. In this study, 95% modeler point coverage was compared to 99% modeler coverage. Two other error sources were investigated: the correlation error and the targeting error. These were added to the prediction error to give an aggregate error for the CyberKnife during treatment of lung tumors. Results: Considering the magnitude of 2{sigma} from the mean of the Gaussian in each signed dimension, the margin expansions needed for 95% modeler point coverage were 1.2 mm in the lateral (LAT) direction and 1.7 mm in the anterior-posterior (AP) direction. For the superior-inferior (SI) direction, the fit was poor; but empirically, the expansions were 3.5 mm. For 99% modeler point coverage, the AP margin was 3.6 mm and the lateral margin was 2.9 mm. The SI margins for 99% modeler point coverage were highly variable. The aggregate error at 95% was 6.9 mm in the SI direction, 4.6 mm in the AP direction, and 3.5 in the lateral direction. Conclusions: The Predictor points follow the Modeler points closely. Margins were found in each clinical direction that would provide 95% modeler point coverage for 95% of the models reviewed in this study. Similar margins were found in two clinical directions for 99% modeler point coverage in 95% of models. These results can offer guidance in the selection of CTV margins for treatment with the CyberKnife.

61 RADIATION PROTECTION AND DOSIMETRY↗

DOE OSTI2015

SU-E-J-254: Evaluating the Role of Mid-Treatment and Post-Treatment FDG-PET/CT in Predicting Progression-Free Survival and Distant Metastasis of Anal Cancer Patients Treated with Chemoradiotherapy

Purpose: To evaluate the role of mid-treatment and post-treatment FDG-PET/CT in predicting progression-free survival (PFS) and distant metastasis (DM) of anal cancer patients treated with chemoradiotherapy (CRT). Methods: 17 anal cancer patients treated with CRT were retrospectively studied. The median prescription dose was 56 Gy (range, 50–62.5 Gy). All patients underwent FDG-PET/CT scans before and after CRT. 16 of the 17 patients had an additional FDG-PET/CT image at 3–5 weeks into the treatment (denoted as mid-treatment FDG-PET/CT). 750 features were extracted from these three sets of scans, which included both traditional PET/CT measures (SUVmax, SUVpeak, tumor diameters, etc.) and spatialtemporal PET/CT features (comprehensively quantify a tumor’s FDG uptake intensity and distribution, spatial variation (texture), geometric property and their temporal changes relative to baseline). 26 clinical parameters (age, gender, TNM stage, histology, GTV dose, etc.) were also analyzed. Advanced analytics including methods to select an optimal set of predictors and a model selection engine, which identifies the most accurate machine learning algorithm for predictive analysis was developed. Results: Comparing baseline + mid-treatment PET/CT set to baseline + posttreatment PET/CT set, 14 predictors were selected from each feature group. Same three clinical parameters (tumor size, T stage and whether 5-FU was held during any cycle of chemotherapy) and two traditional measures (pre- CRT SUVmin and SUVmedian) were selected by both predictor groups. Different mix of spatial-temporal PET/CT features was selected. Using the 14 predictors and Naive Bayes, mid-treatment PET/CT set achieved 87.5% accuracy (2 PFS patients misclassified, all local recurrence and DM patients correctly classified). Post-treatment PET/CT set achieved 94.0% accuracy (all PFS and DM patients correctly predicted, 1 local recurrence patient misclassified) with logistic regression, neural network or support vector machine model. Conclusion: Applying radiomics approach to either midtreatment or post-treatment PET/CT could achieve high accuracy in predicting anal cancer treatment outcomes. This work was supported in part by the National Cancer Institute Grant R01CA172638.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2015

SU-E-J-270: Repeated 18F-FDG PET/CTs Based Feature Analysis for the Predication of Anal Cancer Recurrence

Purpose: To identify PET/CT based imaging predictors of anal cancer recurrence and evaluate baseline vs. mid-treatment vs. post-treatment PET/CT scans in the tumor recurrence prediction. Methods: FDG-PET/CT scans were obtained at baseline, during chemoradiotherapy (CRT, midtreatment), and after CRT (post-treatment) in 17 patients of anal cancer. Four patients had tumor recurrence. For each patient, the mid-treatment and post-treatment scans were respectively aligned to the baseline scan by a rigid registration followed by a deformable registration. PET/CT image features were computed within the manually delineated tumor volume of each scan to characterize the intensity histogram, spatial patterns (texture), and shape of the tumors, as well as the changes of these features resulting from CRT. A total of 335 image features were extracted. An Exact Logistic Regression model was employed to analyze these PET/CT image features in order to identify potential predictors for tumor recurrence. Results: Eleven potential predictors of cancer recurrence were identified with p < 0.10, including five shape features, five statistical texture features, and one CT intensity histogram feature. Six features were indentified from posttreatment scans, 3 from mid-treatment scans, and 2 from baseline scans. These features indicated that there were differences in shape, intensity, and spatial pattern between tumors with and without recurrence. Recurrent tumors tended to have more compact shape (higher roundness and lower elongation) and larger intensity difference between baseline and follow-up scans, compared to non-recurrent tumors. Conclusion: PET/CT based anal cancer recurrence predictors were identified. The post-CRT PET/CT is the most important scan for the prediction of cancer recurrence. The baseline and mid-CRT PET/CT also showed value in the prediction and would be more useful for the predication of tumor recurrence in early stage of CRT. This work was supported in part by the National Cancer Institute Grant R01CA172638.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2019

Prediction of Tropical Cyclone Genesis from Mesoscale Convective Systems using Machine Learning

Tropical cyclone (TC) genesis is a problem of great significance in climate and weather research. Although various environmental conditions necessary for TC genesis have been recognized for a long time, prediction of TC genesis remains a challenge due to complex and stochastic processes involved during TC genesis. Different from traditional statistical and dynamical modeling of TC genesis, in this research, a machine learning framework is developed to determine whether a mesoscale convective system (MCS) would evolve into a tropical cyclone. The machine learning models 1) are built upon a number of essential environmental predictors associated with MCS/TC, 2) predict whether MCSs can become TCs at different lead times, 3) provide information about the relative importance of each predictor, which can be conducive to discovering new aspects of TC genesis. The results indicate that the machine learning classifier, AdaBoost, is able to achieve a 97.2% F1-score accuracy in predicting TC genesis over the entire tropics at a 6-hour lead time using a comprehensive set of environmental predictors. A robust performance can still be attained when the lead time is extended to 12 hours, 24 hours, and 48 hours, and when this machine learning classifier is separately applied to the North Atlantic Ocean and the Western North Pacific Ocean. In contrast, the conventional approach based on genesis potential index can only have no more than 80% F1-score accuracy. Moreover, the machine learning classifier suggests that the low-level vorticity and genesis potential index are the most important predictors to TC genesis, which is consistent with previous discoveries.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2023

Data for The Role of Social Support on Midwestern Farmers’ Willingness to Grow Perennial Bioenergy Crops

The lack of farmers’ willingness to grow perennial bioenergy crops (PBCs) presents a critical barrier to the emergence of cellulosic biofuel production. The willingness relies on a complex network of economic, environmental, and social drivers, among which the influence of social factors (e.g., the influence of neighborhood, community, and communication) is less understood. This study addresses this knowledge gap via a survey analysis of midwestern farmers. The survey data are analyzed through ordinary least square regression and structural equation model, which together investigate the individual and interactive impacts of multiple factors on farmers’ decisions to adopt PBCs. Based on a farm-scale analysis, six statistically significant predictors of farmer willingness to grow PBCs are identified: perception of PBCs’ environment benefits, education level, willingness to take risks, familiarity with PBCs, portion of peers already growing PBCs, and support of biorefineries locating in the local community. Among these, the latter three predictors are social support variables. It is found that familiarity with the crops is the most significant predictor of willingness; familiarity is also an important intermediate variable that mediates the influence of many other predictors. In addition, peer adoption can both directly and indirectly affect willingness via its influence on familiarity. These findings suggest that it is a pressing need to improve farmers’ knowledge of PBCs to promote the adoption of such crops.

Economics↗

DOE OSTI2020

Using machine learning to derive cloud condensation nuclei number concentrations from commonly available measurements

Cloud condensation nuclei (CCN) number concentrations are an important aspect of aerosol–cloud interactions and the subsequent climate effects; however, their measurements are very limited. We use a machine learning tool, random decision forests, to develop a random forest regression model (RFRM) to derive CCN at 0.4 % supersaturation ([CCN0.4]) from commonly available measurements. The RFRM is trained on the long-term simulations in a global size-resolved particle microphysics model. Using atmospheric state and composition variables as predictors, through associations of their variabilities, the RFRM is able to learn the underlying dependence of [CCN0.4] on these predictors, which are as follows: eight fractions of PM 2.5 (NH 4 , SO 4 , NO 3, secondary organic aerosol (SOA), black carbon (BC), primary organic carbon (POC), dust, and salt), seven gaseous species (NO x , NH 3 , O 3 , SO 2 , OH, isoprene, and monoterpene), and four meteorological variables (temperature (T), relative humidity (RH), precipitation, and solar radiation). The RFRM is highly robust: it has a median mean fractional bias (MFB) of 4.4 % with ≈96.33 % of the derived [CCN0.4] within a good agreement range of -60% 2.5 speciation (NH 4 , SO 4 , NO 3 , and organic carbon (OC)), NO x , O 3 , SO 2 , T, and RH, as well as [CCN0.4] are available. We modify, optimize, and retrain the developed RFRM to make predictions from 19 to 9 of these available predictors. This retrained RFRM (RFRM-ShortVars) shows a reduction in performance due to the unavailability and sparsity of measurements (predictors); it captures the [CCN0.4] variability and magnitude at SGP with ≈67.02 % of the derived values in the good agreement range. This work shows the potential of using the more commonly available measurements of PM 2.5 speciation to alleviate the sparsity of CCN number concentrations' measurements.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS1980

Preliminary evaluation of time and distance spacing cues using a cockpit displayed target

A simulator evaluation was conducted to compare time and distance selfspacing cues wherein a pilot establishes and maintains separation from a cockpit displayed target. Langley's terminal configured vehicle simulator which represents an advanced aircraft employing cathode ray tubes for primary flight displays and highly augmented flight control modes was used. Three tasks, two in trail rendezvous tasks (target straight and level and target turning) and one merge task were utilized. Three pilots flew each task twice, once with a time predictor cue and once with a distance predictor. Both time and distance spacing methods are operationally acceptable to the pilots. A slight preference was indicated by the pilots for the distance based predictor in performing rendezvous and merging tasks. Analysis of the recorded data showed the spacing performance with both predictors was essentially the same.

Kelly, J. R.↗

NASA NTRS1991

Optimization techniques applied to passive measures for in-orbit spacecraft survivability

Spacecraft designers have always been concerned about the effects of meteoroid impacts on mission safety. The engineering solution to this problem has generally been to erect a bumper or shield placed outboard from the spacecraft wall to disrupt/deflect the incoming projectiles. Spacecraft designers have a number of tools at their disposal to aid in the design process. These include hypervelocity impact testing, analytic impact predictors, and hydrodynamic codes. Analytic impact predictors generally provide the best quick-look estimate of design tradeoffs. The most complete way to determine the characteristics of an analytic impact predictor is through optimization of the protective structures design problem formulated with the predictor of interest. Space Station Freedom protective structures design insight is provided through the coupling of design/material requirements, hypervelocity impact phenomenology, meteoroid and space debris environment sensitivities, optimization techniques and operations research strategies, and mission scenarios. Major results are presented.

Mog, Robert A.↗

NASA NTRS1991

Steady-state heat conduction in multilayered composite plates and shells

A study is made of a predictor-corrector procedure for the accurate determination of the temperature and heat flux distributions in thick multilayered composite plates and shells. A linear through-the-thickness temperature distribution is used in the predictor phase. The functional dependence of temperature on the thickness coordinate is then calculated a posteriori and used in the corrector phase. Extensive numerical results are presented for linear steady-state heat conduction problems, showing the effects of variation in the geometric and lamination parameters on the accuracy of the thermal response predictions of the predictor-corrector approach. Both antisymmetrically laminated anisotropic plates and multilayered orthotropic cylinders are considered. The solutions are assumed to be periodic in the surface coordinates. For each problem the standard of comparison is taken to be the analytic three-dimensional solution based on treating each layer as a homogeneous anisotropic medium. The potential of the predictor-corrector approach for predicting the thermal response of multilayered plates and shells with complicated geometry is discussed.

Noor, A. K.↗

NASA NTRS1999

"Novel Techniques in Non-Stationary Analysis of Rotorcraft Vibration Signitures"

This research effort produced new methods to analyze the performance of linear predictors that track non-stationary processes. Specifically, prediction methods have been applied to the vibration pattern of rotorcraft drivetrains. This analysis is part or a larger rotorcraft Health and Usage Monitoring System (HUMS) that can diagnose immediate failures of the subsystems, as indicated by abrupt change in the vibration signature, and prognosticate future health, by examining the vibration patterns against long-term trends. This problem is described by a earlier joint paper co-authored by members of the funding agency and the recipient institutions prior to this grant effort. Specific accomplishments under this grant include the following: (1) Definition of a framework for analysis of non-stationary time-series estimation using the coefficients of an adaptive filter. (2) Description of a novel method of combining short-term predictor error and long-term regression error to analyze the performance of a non-stationary predictor. (3) Formulation of a multi-variate probability density function that quantifies the performance of a adaptive predictor by using the short- and long-term error variables in a Gamma function distribution. and (4) Validation of the mathematical formulations with empirical data from NASA flight tests and simulated data to illustrate the utility beyond the domain of vibrating machinery.

Meng, Teresa↗

NASA NTRS2000

Sleep and Mood During A Winter in Antarctica

Seasonal variations in sleep characteristics and their association with changes in mood were examined in 91 American men and women also who spent the 1991 austral winter at three different research stations in Antarctica. Measures of total hours of sleep over a 24-hr period, duration of longest (i.e.,"nighttime") sleep event, number of sleep events, time of sleep onset, and quality of sleep remained unchanged over the course of the austral winter (March through October). However, exposure to total darkness based on station latitude was significantly associated with total hours of sleep, duration of are longest sleep event, time of sleep onset, and quality of sleep. Reported vigor the previous month was a significant independent predictor of changes in all five sleep measures; previous month's measures of all six POMS subscales were significant independent predictors of sleep quality. Sleep characteristics were significant independent predictors of vigor and confusion the following month; total sleep, longest sleep event, sleep onset and sleep quality were significant independent predictors of tension-anxiety and depression. Changes in mood during the austral winter are preceded by changes in sleep characteristics, but prolonged exposure to the photoperiodicity characteristic of the high latitudes appears to be associated with improved sleep. In turn, mood changes appear to affect certain sleep characteristics, especially sleep quality.

Palinkas, Lawrence A.↗

NASA NTRS2001

A New Ensemble Canonical Correlation Prediction Scheme for Seasonal Precipitation

Department of Mathematical Sciences, University of Alberta, Edmonton, Canada This paper describes the fundamental theory of the ensemble canonical correlation (ECC) algorithm for the seasonal climate forecasting. The algorithm is a statistical regression sch eme based on maximal correlation between the predictor and predictand. The prediction error is estimated by a spectral method using the basis of empirical orthogonal functions. The ECC algorithm treats the predictors and predictands as continuous fields and is an improvement from the traditional canonical correlation prediction. The improvements include the use of area-factor, estimation of prediction error, and the optimal ensemble of multiple forecasts. The ECC is applied to the seasonal forecasting over various parts of the world. The example presented here is for the North America precipitation. The predictor is the sea surface temperature (SST) from different ocean basins. The Climate Prediction Center's reconstructed SST (1951-1999) is used as the predictor's historical data. The optimally interpolated global monthly precipitation is used as the predictand?s historical data. Our forecast experiments show that the ECC algorithm renders very high skill and the optimal ensemble is very important to the high value.

Kim, Kyu-Myong↗

NASA NTRS2000

Fractal correlation properties of R-R interval dynamics and mortality in patients with depressed left ventricular function after an acute myocardial infarction

BACKGROUND: Preliminary data suggest that the analysis of R-R interval variability by fractal analysis methods may provide clinically useful information on patients with heart failure. The purpose of this study was to compare the prognostic power of new fractal and traditional measures of R-R interval variability as predictors of death after acute myocardial infarction. METHODS AND RESULTS: Time and frequency domain heart rate (HR) variability measures, along with short- and long-term correlation (fractal) properties of R-R intervals (exponents alpha(1) and alpha(2)) and power-law scaling of the power spectra (exponent beta), were assessed from 24-hour Holter recordings in 446 survivors of acute myocardial infarction with a depressed left ventricular function (ejection fraction </=35%). During a mean+/-SD follow-up period of 685+/-360 days, 114 patients died (25.6%), with 75 deaths classified as arrhythmic (17.0%) and 28 as nonarrhythmic (6.3%) cardiac deaths. Several traditional and fractal measures of R-R interval variability were significant univariate predictors of all-cause mortality. Reduced short-term scaling exponent alpha(1) was the most powerful R-R interval variability measure as a predictor of all-cause mortality (alpha(1) <0.75, relative risk 3.0, 95% confidence interval 2.5 to 4.2, P<0.001). It remained an independent predictor of death (P<0.001) after adjustment for other postinfarction risk markers, such as age, ejection fraction, NYHA class, and medication. Reduced alpha(1) predicted both arrhythmic death (P<0.001) and nonarrhythmic cardiac death (P<0.001). CONCLUSIONS: Analysis of the fractal characteristics of short-term R-R interval dynamics yields more powerful prognostic information than the traditional measures of HR variability among patients with depressed left ventricular function after an acute myocardial infarction.

Non-NASA Center↗

NASA NTRS1996

Power law behavior of RR-interval variability in healthy middle-aged persons, patients with recent acute myocardial infarction, and patients with heart transplants

BACKGROUND. The purposes of the present study were (1) to establish normal values for the regression of log(power) on log(frequency) for, RR-interval fluctuations in healthy middle-aged persons, (2) to determine the effects of myocardial infarction on the regression of log(power) on log(frequency), (3) to determine the effect of cardiac denervation on the regression of log(power) on log(frequency), and (4) to assess the ability of power law regression parameters to predict death after myocardial infarction. METHODS AND RESULTS. We studied three groups: (1) 715 patients with recent myocardial infarction; (2) 274 healthy persons age and sex matched to the infarct sample; and (3) 19 patients with heart transplants. Twenty-four-hour RR-interval power spectra were computed using fast Fourier transforms and log(power) was regressed on log(frequency) between 10(-4) and 10(-2) Hz. There was a power law relation between log(power) and log(frequency). That is, the function described a descending straight line that had a slope of approximately -1 in healthy subjects. For the myocardial infarction group, the regression line for log(power) on log(frequency) was shifted downward and had a steeper negative slope (-1.15). The transplant (denervated) group showed a larger downward shift in the regression line and a much steeper negative slope (-2.08). The correlation between traditional power spectral bands and slope was weak, and that with log(power) at 10(-4) Hz was only moderate. Slope and log(power) at 10(-4) Hz were used to predict mortality and were compared with the predictive value of traditional power spectral bands. Slope and log(power) at 10(-4) Hz were excellent predictors of all-cause mortality or arrhythmic death. To optimize the prediction of death, we calculated a log(power) intercept that was uncorrelated with the slope of the power law regression line. We found that the combination of slope and zero-correlation log(power) was an outstanding predictor, with a relative risk of > 10, and was better than any combination of the traditional power spectral bands. The combination of slope and log(power) at 10(-4) Hz also was an excellent predictor of death after myocardial infarction. CONCLUSIONS. Myocardial infarction or denervation of the heart causes a steeper slope and decreased height of the power law regression relation between log(power) and log(frequency) of RR-interval fluctuations. Individually and, especially, combined, the power law regression parameters are excellent predictors of death of any cause or arrhythmic death and predict these outcomes better than the traditional power spectral bands.

Non-NASA Center↗

NASA NTRS2007

Advanced Transport Delay Compensation Algorithms: Results of Delay Measurement and Piloted Performance Tests

This report summarizes the results of delay measurement and piloted performance tests that were conducted to assess the effectiveness of the adaptive compensator and the state space compensator for alleviating the phase distortion of transport delay in the visual system in the VMS at the NASA Langley Research Center. Piloted simulation tests were conducted to assess the effectiveness of two novel compensators in comparison to the McFarland predictor and the baseline system with no compensation. Thirteen pilots with heterogeneous flight experience executed straight-in and offset approaches, at various delay configurations, on a flight simulator where different predictors were applied to compensate for transport delay. The glideslope and touchdown errors, power spectral density of the pilot control inputs, NASA Task Load Index, and Cooper-Harper rating of the handling qualities were employed for the analyses. The overall analyses show that the adaptive predictor results in slightly poorer compensation for short added delay (up to 48 ms) and better compensation for long added delay (up to 192 ms) than the McFarland compensator. The analyses also show that the state space predictor is fairly superior for short delay and significantly superior for long delay than the McFarland compensator.

Guo, Liwen↗

NASA NTRS2007

Automated Knowledge Discovery From Simulators

A computational method, SimLearn, has been devised to facilitate efficient knowledge discovery from simulators. Simulators are complex computer programs used in science and engineering to model diverse phenomena such as fluid flow, gravitational interactions, coupled mechanical systems, and nuclear, chemical, and biological processes. SimLearn uses active-learning techniques to efficiently address the "landscape characterization problem." In particular, SimLearn tries to determine which regions in "input space" lead to a given output from the simulator, where "input space" refers to an abstraction of all the variables going into the simulator, e.g., initial conditions, parameters, and interaction equations. Landscape characterization can be viewed as an attempt to invert the forward mapping of the simulator and recover the inputs that produce a particular output. Given that a single simulation run can take days or weeks to complete even on a large computing cluster, SimLearn attempts to reduce costs by reducing the number of simulations needed to effect discoveries. Unlike conventional data-mining methods that are applied to static predefined datasets, SimLearn involves an iterative process in which a most informative dataset is constructed dynamically by using the simulator as an oracle. On each iteration, the algorithm models the knowledge it has gained through previous simulation trials and then chooses which simulation trials to run next. Running these trials through the simulator produces new data in the form of input-output pairs. The overall process is embodied in an algorithm that combines support vector machines (SVMs) with active learning. SVMs use learning from examples (the examples are the input-output pairs generated by running the simulator) and a principle called maximum margin to derive predictors that generalize well to new inputs. In SimLearn, the SVM plays the role of modeling the knowledge that has been gained through previous simulation trials. Active learning is used to determine which new input points would be most informative if their output were known. The selected input points are run through the simulator to generate new information that can be used to refine the SVM. The process is then repeated. SimLearn carefully balances exploration (semi-randomly searching around the input space) versus exploitation (using the current state of knowledge to conduct a tightly focused search). During each iteration, SimLearn uses not one, but an ensemble of SVMs. Each SVM in the ensemble is characterized by different hyper-parameters that control various aspects of the learned predictor - for example, whether the predictor is constrained to be very smooth (nearby points in input space lead to similar output predictions) or whether the predictor is allowed to be "bumpy." The various SVMs will have different preferences about which input points they would like to run through the simulator next. SimLearn includes a formal mechanism for balancing the ensemble SVM preferences so that a single choice can be made for the next set of trials.

Burl, Michael↗

NASA NTRS2010

Analysis of Flight Management System Predictions of Idle-Thrust Descents

To enable arriving aircraft to fly optimized descents computed by the flight management system (FMS) in congested airspace, ground automation must accurately predict descent trajectories. To support development of the predictor and its uncertainty models, descents from cruise to the meter fix were executed using vertical navigation in a B737-700 simulator and a B777-200 simulator, both with commercial FMSs. For both aircraft types, the FMS computed the intended descent path for a specified speed profile assuming idle thrust after top of descent (TOD), and then it controlled the avionics without human intervention. The test matrix varied aircraft weight, descent speed, and wind conditions. The first analysis in this paper determined the effect of the test matrix parameters on the FMS computation of TOD location, and it compared the results to those for the current ground predictor in the Efficient Descent Advisor (EDA). The second analysis was similar but considered the time to fly a specified distance to the meter fix. The effects of the test matrix variables together with the accuracy requirements for the predictor will determine the allowable error for the predictor inputs. For the B737, the EDA prediction of meter fix crossing time agreed well with the FMS; but its prediction of TOD location probably was not sufficiently accurate to enable idle-thrust descents in congested airspace, even though the FMS and EDA gave similar shapes for TOD location as a function of the test matrix variables. For the B777, the FMS and EDA gave different shapes for the TOD location function, and the EDA prediction of the TOD location is not accurate enough to fully enable the concept. Furthermore, the differences between the FMS and EDA predictions of meter fix crossing time for the B777 indicated that at least one of them was not sufficiently accurate.

Stell, Laurel↗