Search NASA⌕ Search

SEARCH · Search NASA

Results for “input combinations”

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.

At least 163 records · Page 9

Building a Real-Time Flood Prediction Model for Improving Early Warning Systems in Ellicott City, Maryland

As flood events in the United States grow in frequency and intensity, the use of applied remote sensing analyses is increasingly necessary for effective flood monitoring and warning systems. The NASA DEVELOP National Program partnered with the local government of Howard County, Maryland, to investigate the use of machine learning for advanced flood risk detection, and to test the feasibility of integrating this approach into the county’s flood early warning system. To strengthen the efforts of the Howard County Office of Emergency Management (OEM), the project developed a statistical model capable of hindcasting the two severe flash flood events that devastated Ellicott City and transitioned to a ‘Long Short-Term Memory’ based sequence-to-sequence deep learning model with 8-hour forecast capability. The team combined data inputs from public sources including river and precipitation gauges, NASA and NOAA Earth observations, and numerical weather model products using scripts written in the Google Colaboratory Python scripting environment. In addition to designing the deep learning architecture, the team trained and tested the model, and evaluated its performance using Nash-Sutcliffe Efficiency. The final product, the Sequentially Trained Real-time EstimAted Model (STREAM) predicts stage height for the Hudson Branch gauge in Ellicott City using data products available in near real-time, including the High-Resolution Rapid Refresh model’s accumulated precipitation forecasts supplemented by stream gauge data from the OEM and the U.S. Geological Survey. STREAM was incorporated into an online dashboard in a user-friendly interface capable of triggering the alarms that initiate emergency response protocols up to 8 hours in advance of a predicted severe flood event. The project demonstrated the potential for the integration of open data and Earth observations into a flood risk forecasting tool capable of informing near real-time decision making.

NASA DEVELOP↗

Building a Real-Time Predictive Flood Model for Improving Early Warning Systems in Ellicott City, Maryland

As flood events in the United States grow in frequency and intensity, the use of applied remote sensing analyses is increasingly necessary for effective flood monitoring and warning systems. The NASA DEVELOP National Program partnered with the Howard County government in Maryland to investigate the use of machine learning for advanced flood risk detection, and to test the feasibility of integrating this approach into the county’s flood early warning system. To strengthen the efforts of the Howard County Office of Emergency Management (OEM), the project developed a prediction model capable of hindcasting the two severe flash flood events that devastated Ellicott City, and transitioned to an Long Short-Term Memory (LSTM) based sequence-to-sequence deep learning model with 8-hour forecast capability. The team combined data inputs from public sources including river and precipitation gauges, NASA and NOAA Earth observations, and numerical weather model products using scripts written in the Google Colaboratory Python scripting environment. In addition to designing the deep learning architecture, the team trained and tested the model, and evaluated its performance using the Nash-Sutcliffe Efficiency (NSE). The final product, called the Sequentially Trained Real-time EstimAted Model (STREAM), predicts stage height for the Hudson Branch gauge in Ellicott City using data products available in near real-time, including the High-Resolution Rapid Refresh (HRRR) model’s accumulated precipitation forecasts supplemented by stream gauge data from the OEM and the U.S. Geological Survey. STREAM was incorporated into an online dashboard in a user-friendly interface capable of triggering the alarms that initiate the OEM’s emergency response protocols up to 8 hours in advance of a predicted severe flood event. The project demonstrated the potential for the integration of open data and Earth observations into a flood risk forecasting tool capable of informing near real-time decision making.

Ryan Hammock↗

Effects of breach formation parameter uncertainty on inundation risk area and consequence analysis

According to the national inventory of dams (NID), there are approximately 79,500 dams in the United States, with 11,800 of these dams being classified as high-hazard. It has been recommended that each high-hazard dam in the United States have an emergency action plan (EAP), but it has been found that only about 60% of the high-hazard dams have a complete EAP. A major aspect of these plans is inundation risk area identification and associated impacts in the event of dam failure. In order to determine the inundation risk area an estimation of breach discharge must be completed. Most methods used to determine breach discharge, including the NWS-DAMBRK model, require modelers to select size, shape, and time of breach formation. Federal agencies (e.g. Bureau of Reclamation, Federal Energy Regulatory Commission) with oversight of U.S. dams have recommended ranges of values for each of these parameters based on dam type. However, variations in these parameters even within the recommended range have the potential to impose significant transformation on the discharge hydrograph relative to both timing and magnitude of the peak discharge. Therefore, it has also been recommended that sensitivity of these parameters be investigated when performing breach inundation analyses. This paper presents a sensitivity analysis of three breach parameters (average breach width, side slope, and time to failure) on a case study dam located in the United States. The sensitivity analysis employed was based on the 3{sup 3} factorial design, in which three levels (e.g. low, medium, and high) were selected for each of the three parameters, resulting in twenty-seven combinations. The three levels remained within the recommended range of values for each parameter type. With each combination of input parameters, a discharge hydrograph was generated and used as a source condition for inundation analysis using a two-dimensional shallow water equation model. The resulting simulations were compared to determine the sensitivity of flood inundation area, flood arrival time, peak flood depths, and socio-economic impacts (e.g. population at risk, direct and indirect economic loss) to changes in individual parameters and parameter interactions. Results and discussion from this sensitivity analysis will be presented in detail in the paper.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Physics-informed Machine Learning Method for Forecasting and Uncertainty Quantification of Partially Observed and Unobserved States in Power Grids

We present a physics-informed Gaussian Process Regression (GPR) model to predict the phase angle, angular speed, and wind mechanical power from a limited number of measurements. In the traditional data-driven GPR method, the form of the Gaussian Process covariance matrix is assumed and its parameters are found from measurements. In the physics-informed GPR, we treat unknown variables (including wind speed and mechanical power) as a random process and compute the covariance matrix from the resulting stochastic power grid equations. We demonstrate that the physics-informed GPR method is significantly more accurate than the standard data-driven one for immediate forecasting of generators' angular velocity and phase angle. We also show that the physics-informed GPR provides accurate predictions of the unobserved wind mechanical power, phase angle, or angular velocity when measurements from only one of these variables are available. The immediate forecast of observed variables and predictions of unobserved variables can be used for effectively managing power grids (electricity market clearing, regulation actions) and early detection of abnormal behavior and faults. The physics-based GPR forecast time horizon depends on the combination of input (wind power, load, etc.) correlation time and characteristic (relaxation) time of the power grid and can be extended to short and medium-range times.

machine learning, power grid, forecasting, Gaussia↗

Multimodal Data Representation with Deep Learning for Extracting Cancer Characteristics from Clinical Text

This paper presents a multimodal data representation to improve the performance of deep learning models for extracting cancer key characteristics from unstructured text in pathology reports. Specifically, in addition to using the text as the input to deep learning models, we use concept unique identifiers (CUIs) as another source of information to the models. We analyze the performance of different text and CUI data representations, including word embeddings and bag of embeddings (BOE), with a convolutional neural network (CNN) and a fully connected multilayer perceptron neural network (MLP-NN). The high level document embeddings from text and CUI inputs are combined by concatenating them and then applying a classifier. The model is used for extracting cancer subsite and histology from pathology reports. These two classification tasks have a large number of labels, i.e. 317 for subsite and 556 for histology, with extreme class imbalance. We compare the performance of the developed DL models across the two tasks based on micro- and macro-F1 scores. The evaluation shows that a multi-channel DL model that utilizes text represented by word embeddings and CUIs represented by BOE outperforms other DL models. Also, this approach significantly improves the model performance on low prevalence classes.

Alawad, Mohammed↗

Simple Building Calculator

Whole building energy modeling is a powerful tool for analyzing energy use in buildings. For large buildings benefits of modeling services can easily be quantified due to the short payback period of the implemented measures. For small buildings, the upfront costs of developing robust energy models often deter building owners from investing in energy modeling services. To address this shortcoming, a ”Simple Building Calculator” (SBC) was developed. SBC uses pre-simulated results of a range of common measures over a wide range of efficiency inputs. It combines whole building simulation results with statistical modeling techniques to predict energy impact of measures very quickly. In this paper we present the modeling methodology used to develop the data supporting SBC.

Building Energy Simulation, commercial building, E↗

Remediation of acid mine drainage with sulfate reducing bacteria

Sulfate reducing bacteria have been shown to be effective at treating acid mine drainage through sulfide production and subsequent precipitation of metal sulfides. In this laboratory experiment for undergraduate environmental chemistry courses, students design and implement a set of bioreactors to remediate acid mine drainage and explain observed changes in dissolved metal concentrations and pH. Using synthetic acid mine drainage and combinations of inputs, students monitor their bioreactors for decreases in dissolved copper and iron concentrations.

01 COAL, LIGNITE, AND PEAT↗

Radiatively important parameters best estimate (RIPBE) value-added product (VAP)

Currently, to calculate radiative heating rate profiles for the Broadband Heating Rate Profile (BBHRP) product, radiatively important parameters (water vapor, ozone, surface albedo, aerosol properties, and cloud properties) from multiple VAPs and datastreams are combined into input text files that are then used to run the RRTM radiative transfer codes. These input parameters have different temporal and spatial scales and are difficult to extract from the text files to be used for other purposes such as running other radiative transfer codes, analyzing results, or error tracking. The purpose of the Radiatively Important Parameters Best Estimate (RIPBE) VAP is to improve this process by creating a clearly identified set of inputs for BBHRP (and other radiation codes) on a uniform vertical and temporal grid. This process will decouple the input/output portion of the BBHRP from the core physics (the RRTM radiative transfer model) and will add error tracking and version information to the input data set. Critical parameters (which must exist for the radiation code to be run) will be designated; for other parameters, climatological or fixed values will be used when the preferred values are missing. This should increase the number of cases for which radiative transfer calculations can be run. In all cases, flags will clearly identify the source for each parameter. RIPBE will serve multiple functions: (1) it will provide a clearly identifiable set of inputs for BBHRP, (2) it will facilitate the use of BBHRP as a retrieval and radiation code development testbed by providing a vehicle for easily extracting and swapping input parameters needed to conduct radiative transfer calculations, and (3) it will be a complement to the Climate Modeling Best Estimate (CMBE) VAP and will provide a significantly expanded set of parameters for model evaluation in a showcase data set form. At the ASR meeting, we will present examples and evaluation of the initial RIPBE dataset at SGP.

54 ENVIRONMENTAL SCIENCES↗

Model Inputs, Outputs, and Scripts associated with: “Combined effects of stream hydrology and land use on basin-scale hyporheic zone denitrification in the Columbia River Basin”

This data package is associated with the publication “Combined effects of stream hydrology and land use on basin‐scale hyporheic zone denitrification in the Columbia River Basin”, published in Water Resource Research (Son et al.2022) available at https://doi.org/10.1029/2021WR031131. This data package includes the key model inputs/outputs of the river corridor model for the Columbia River Basin (CRB) and the model source codes used in the manuscript. The model is a carbon-nitrogen-coupled river corridor model (RCM), and the model is used to quantify hyporheic zone (HZ) denitrification at the NHDPLUS stream reach scales. The RCM used in this study combines empirical substrate models derived from observations and three microbially driven reactions, including two-step denitrification and aerobic respiration, are considered within the HZ. The key input data of the model are exchange flux, residence time, and stream solute (dissolved organic carbon (DOC), dissolved oxygen (DO), and nitrate concentrations). These inputs are constant over time and represent long-term averaged values. This study uses the RCM to explore the spatial patterns of HZ denitrification across reaches with different sizes and land use in the CRB. Our main objective is to use the RCM as a virtual reality model, and the machine-learning models as surrogates that encapsulate the complexities of the physics-based model while identifying the importance of different variables that are not evident in the model conceptualization. We do not include a direct comparison of the modeled HZ denitrification and measurements; however, the RCM can capture the overall spatial patterns of the HZ denitrification because the model inputs and its reaction networks are based on well-established theory and a physical-based model. The combination of the model-based predictions and a machine-learning approach (e.g., random forest) is used to improve our understanding of what variables of the model are associated with spatial patterns of the modeled denitrification across reaches with different sizes and land uses, and to develop a proxy model using measurable variables to reproduce the simulated patterns.This dataset contains five folders: (1) model_inputs, (2) model_outputs, (3) Rscripts, (4) figures, and (5) model_codes. It also contains a readme, file level metadata (FLMD), and data dictionary (dd). Please see the FLMD for a list of all the files contained in this data package and descriptions for each. The model_inputs folder contains the model inputs used to drive the model simulations. The model_outputs folder contains key model output files from the river corridor model. The Rscripts folder contains the Rscripts for pre- and post- processing model results. The figures folder contains the raw figures associated with the manuscript. The model_codes folder includes key model source codes/input files. All files are .jpg, .jpeg, .out, .e, .od, .dat, .sub, .F90, .0, .R, .sbx, .cpg, .sbn, .shx, .shp, .dbf, .prj, .tfw, .tif, .xml, .pdf, or .csv.

54 ENVIRONMENTAL SCIENCES↗

Weldability of irradiated Stainless Steel 304 materials harvested from the National Research Universal (NRU) reactor

In this work, the weldability of the neutron-irradiated Stainless Steel 304 (SS 304) materials containing estimated helium levels of 12 to 44 appm has been studied through a Collaborative Research and Development Agreement between the Canadian Nuclear Laboratories and Oak Ridge National Laboratory. The work utilizes a SS 304 component harvested from the National Research Universal reactor. Laser beam welding was used to explore suitable welding conditions and associated parameters, to determine helium concentration limits for successful welding of irradiated stainless-steels. The experimental results show that maximum fusion zone void sizes, fusion zone void quantity and total heat affected zone helium-induced crack length generally increased with increasing He levels and effective weld heat input. A proper combination of welding parameters with low effective heat input, such as, laser power, weld travel speed and wire feed speed, were shown to improve the welding quality, and potentially reduce the length and probability of the formation of grain boundary helium-induced cracking and voids of irradiated SS 304 even with He concentrations as high as about 44 appm.

36 MATERIALS SCIENCE↗

Uncertainty Analysis of RELAP5-3D©

As world-wide energy consumption continues to increase, so does the demand for the use of alternative energy sources, such as Nuclear Energy. Nuclear Power Plants currently supply over 370 gigawatts of electricity, and more than 60 new nuclear reactors have been commissioned by 15 different countries. The primary concern for Nuclear Power Plant operation and lisencing has been safety. The safety of the operation of Nuclear Power Plants is no simple matter- it involves the training of operators, design of the reactor, as well as equipment and design upgrades throughout the lifetime of the reactor, etc. To safely design, operate, and understand nuclear power plants, industry and government alike have relied upon the use of best-estimate simulation codes, which allow for an accurate model of any given plant to be created with well-defined margins of safety. The most widely used of these best-estimate simulation codes in the Nuclear Power industry is RELAP5-3D. Our project focused on improving the modeling capabilities of RELAP5-3D by developing uncertainty estimates for its calculations. This work involved analyzing high, medium, and low ranked phenomena from an INL PIRT on a small break Loss-Of-Coolant Accident as wall as an analysis of a large break Loss-Of- Coolant Accident. Statistical analyses were performed using correlation coefficients. To perform the studies, computer programs were written that modify a template RELAP5 input deck to produce one deck for each combination of key input parameters. Python scripting enabled the running of the generated input files with RELAP5-3D on INL’s massively parallel cluster system. Data from the studies was collected and analyzed with SAS. A summary of the results of our studies are presented.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DEEPEN 3D PFA Index Models for Exploration Datasets at Newberry Volcano

DEEPEN stands for DE-risking Exploration of geothermal Plays in magmatic ENvironments. As part of the development of the DEEPEN 3D play fairway analysis (PFA) methodology for magmatic plays (conventional hydrothermal, superhot EGS, and supercritical), index models needed to be developed to map values in geoscientific exploration datasets to favorability index values. This GDR submission includes those index models. Index models were created by binning values in exploration datasets into chunks based on their favorability, and then applying a number between 0 and 5 to each chunk, where 0 represents very unfavorable data values and 5 represents very favorable data values. To account for differences in how exploration methods are used to detect each play component, separate index models are produced for each exploration method for each component of each play type. Index models were created using histograms of the distributions of each exploration dataset in combination with literature and input from experts about what combinations of geophysical, geological, and geochemical signatures are considered favorable at Newberry. This is in attempt to create similar sized bins based on the current understanding of how different anomalies map to favorable areas for the different types of geothermal plays (i.e., conventional hydrothermal, superhot EGS, and supercritical). For example, an area of partial melt would likely appear as an area of low density, high conductivity, low vp, and high vp/vs. This means that these target anomalies would be given high (4 or 5) index values for the purpose of imaging the heat source. To account for differences in how exploration methods are used to detect each play component, separate index models are produced for each exploration method for each component of each play type. Index models were produced for the following datasets: - Geologic model - Alteration model - vp/vs - vp - vs - Temperature model - Seismicity (density*magnitude) - Density - Resistivity - Fault distance - Earthquake cutoff depth model

15 GEOTHERMAL ENERGY↗

Modeling of Coronal Mass Ejections that Caused Particularly Large Geomagnetic Storms Using ENLIL Heliosphere Cone Model

In our previous paper we reported the results of modeling of 14 selected well -observed strong halo coronal mass ejection (CME) events using the WSA -ENLIL cone model combination. Cone model input parameters were obtained from white light coronagraph images of the CME events using the analytical method developed by Xie et al. This work verified that coronagraph input gives reasonably good results for the CME arrival time prediction. In contrast to Taktakishvili et al., where we started the analysis by looking for clear CME signatures in the data and then proceeded to model the interplanetary consequences at 1 AU, in the present paper we start by generating a list of observed geomagnetic storm events and then work our way back to remote solar observations and carry out the corresponding CME modeling. The approach used in this study is addressing space weather forecasting and operational needs. We analyzed 36 particularly strong geomagnetic storms, then tried to associate them with particular CMEs using SOHO/LASCO catalogue, and finally modeled these CMEs using WSA-ENLIL cone model. Recently, Pulkkinen et al. developed a novel method for automatic determination of cone model parameters. We employed both analytical and automatic methods to determine cone model input parameters. We examined the CME arrival times and magnitude of impact at 1 AU for both techniques. The results of the simulations are compared with the ACE satellite observations. This comparison demonstrated that WSA -ENLIL model combination with coronagraph input gives reasonably good results for the CME arrival times for this set of 'geoeffective" CME events as well.

Taktakishvili, A.↗

Optically controllable molecular logic circuits

Molecular logic circuits represent a promising technology for observation and manipulation of biological systems at the molecular level. However, the implementation of molecular logic circuits for temporal and programmable operation remains challenging. In this paper, we demonstrate an optically controllable logic circuit that uses fluorescence resonance energy transfer (FRET) for signaling. The FRET-based signaling process is modulated by both molecular and optical inputs. Based on the distance dependence of FRET, the FRET pathways required to execute molecular logic operations are formed on a DNA nanostructure as a circuit based on its molecular inputs. In addition, the FRET pathways on the DNA nanostructure are controlled optically, using photoswitching fluorescent molecules to instruct the execution of the desired operation and the related timings. The behavior of the circuit can thus be controlled using external optical signals. As an example, a molecular logic circuit capable of executing two different logic operations was studied. The circuit contains functional DNAs and a DNA scaffold to construct two FRET routes for executing Input 1 AND Input 2 and Input 1 AND NOT Input 3 operations on molecular inputs. The circuit produced the correct outputs with all possible combinations of the inputs by following the light signals. Moreover, the operation execution timings were controlled based on light irradiation and the circuit responded to time-dependent inputs. The experimental results demonstrate that the circuit changes the output for the required operations following the input of temporal light signals.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A Ka-Band Wide-Bandgap Solid-State Power Amplifier: Architecture Performance Estimates

Motivated by recent advances in wide-bandgap (WBG) gallium nitride (GaN) semiconductor technology, there is considerable interest in developing efficient solidstate power amplifiers (SSPAs) as an alternative to the traveling-wave tube amplifier (TWTA) for space applications. This article documents the results of a study to investigate power-combining technology and SSPA architectures that can enable a 120-W, 40 percent power-added efficiency (PAE) SSPA. Results of the study indicate that architectures based on at least three power combiner designs are likely to enable the target SSPA. The proposed architectures can power combine 16 to 32 individual monolithic microwave integrated circuits (MMICs) with >80 percent combining efficiency. This corresponds to MMIC requirements of 5- to 10-W output power and >48 percent PAE. For the three proposed architectures [1], detailed analysis and design of the power combiner are presented. The first architecture studied is based on a 16-way septum combiner that offers low loss and high isolation over the design band of 31 to 36 GHz. Analysis of a 2-way prototype septum combiner had an input match >25 dB, output match >30 dB, insertion loss <0.02 dB, and isolation >30 dB over the design band. A 16-way design, based on cascading this combiner in a binary fashion, is documented. The second architecture is based on a 24-way waveguide radial combiner. A prototype 24-way radial base was analyzed to have an input match >30 dB (under equal excitation of all input ports). The match of the mode transducer that forms the output of a radial combiner was found to be >27 dB. The functional bandwidth of the radial base and mode transducer, which together will form a radial combiner/divider, exceeded the design band. The third architecture employs a 32-way, parallel-plate radial combiner. Simulation results indicated an input match >24 dB, output match >22 dB, insertion loss <0.23 dB, and adjacent port isolation >20 dB over the design band. All three architectures utilize a low-loss MMIC amplifier module based on commercial MMIC packaging and a custom microstrip-to-rectangular-waveguide transition. The insertion loss of the module is expected to be 0.45 dB over the design band.

Epp, L.↗

Moirai Land Data System (Moirai) v3.0.0

The Moirai Land Data System (Moirai LDS) is designed to produce recent historical land data inputs for the AgLU module of GCAM data system , but the Moirai LDS outputs could also be used by other models/applications. The Moirai are the Greek Fates, and this software is named Moirai to represent the fundamental influence of land data inputs on model outcomes. The primary function of the Moirai LDS is to combine spatially explicit input data (e.g., raster images) with tabular input data (e.g., crop price table) to generate tabular output data for a suite of variables. Some of these outputs replace the data provide by the Global Trade Analysis Project (GTAP), and other data replace and augment the original GCAM GIS processing. The Moirai LDS output data are aggregated by Geographic Land Unit (GLU) within each country. The GLU coverage is an input to the Moirai LDS (as a thematic raster image and an associated CSV file that maps the thematic integers to names), and the GLU boundaries can be determined arbitrarily. Previous versions of GCAM (and Moirai LDS) used only bioclimatic Agro-Ecological Zones (AEZs) and corresponding data that were provided by GTAP, as the GLUs. As a result, some AEZ terminology still exists in the code, but this terminology now refers more generally to GLUs. The Moirai LDS now enables any set of boundaries to be used as GLUs (including AEZs), allowing for more flexible generation of land use region boundaries (defined as the intersection of GLUs with geopolitical regions). The current default set of GLUs is the same set of 235 global watersheds as used by the GCAM water module. The GCAM 5.1 geopolitical regions (32 or 14) are included and used as inputs to Moirai to generate a mapping file between the Moirai outputs, which are at the level of the intersection between the GLUs and the country boundaries, and the geopolitical regions. The diagnostics scripts use this geopolitical region mapping in some cases. Moirai can also recalibrate three of the outputs (crop production, harvested area, and land rent) to a specified year that is the center of a five-year averaging window. No recalibration retains the circa 2000, 7-year average of the source data. The cucrrency-year for land rent can also be specified, and the default is 2001 to match the GTAP data.

Di Vittorio, Alan↗

SMART (Sandia's Modular Architecture for Robotics and Teleoperation) Ver. 1.0

"SMART Ver. 0.8 Beta" provides a system developer with software tools to create a telerobotic control system, i.e., a system whereby an end-user can interact with mechatronic equipment. It consists of three main components: the SMART Editor (tsmed), the SMART Real-time kernel (rtos), and the SMART Supervisor (gui). The SMART Editor is a graphical icon-based code generation tool for creating end-user systems, given descriptions of SMART modules. The SMART real-time kernel implements behaviors that combine modules representing input devices, sensors, constraints, filters, and robotic devices. Included with this software release is a number of core modules, which can be combined with additional project and device specific modules to create a telerobotic controller. The SMART Supervisor is a graphical front-end for running a SMART system. It is an optional component of the SMART Environment and utilizes the TeVTk windowing and scripting environment. Although the code contained within this release is complete, and can be utilized for defining, running, and interfacing to a sample end-user SMART system, most systems will include additional project and hardware specific modules developed either by the system developer or obtained independently from a SMART module developer. SMART is a software system designed to integrate the different robots, input devices, sensors and dynamic elements required for advanced modes of telerobotic control. "SMART Ver. 0.8 Beta" defines and implements a telerobotic controller. A telerobotic system consists of combinations of modules that implement behaviors. Each real-time module represents an input device, robot device, sensor, constraint, connection or filter. The underlying theory utilizes non-linear discretized multidimensional network elements to model each individual module, and guarantees that upon a valid connection, the resulting system will perform in a stable fashion. Different combinations of modules implement different behaviors. Each module must have at a minimum an initialization routine, a parameter adjustment routine, and an update routine. The SMART runtime kernel runs continuously within a real-time embedded system. Each module is first set-up by the kernel, initialized, and then updated at a fixed rate whenever it is in context. The kernel responds to operator directed commands by changing the state of the system, changing parameters on individual modules, and switching behavioral modes. The SMART Editor is a tool used to define, verify, configure and generate source code for a SMART control system. It uses icon representations of the modules, code patches from valid configurations of the modules, and configuration files describing how a module can be connected into a system to lead the end-user in through the steps needed to create a final system. The SMART Supervisor serves as an interface to a SMART run-time system. It provides an interface on a host computer that connects to the embedded system via TCPIIP ASCII commands. It utilizes a scripting language (Tel) and a graphics windowing environment (Tk). This system can either be customized to fit an end-user's needs or completely replaced as needed.

Anderson, Robert↗

Latent map Gaussian processes for mixed variable metamodeling

Gaussian processes (GPs) are ubiquitously used in sciences and engineering as metamodels. Standard GPs, however, can only handle numerical or quantitative variables. Here we introduce latent map Gaussian processes (LMGPs) that inherit the attractive properties of GPs and are also applicable to mixed data which have both quantitative and qualitative inputs. The core idea behind LMGPs is to learn a continuous, low-dimensional latent space or manifold which encodes all qualitative inputs. To learn this manifold, we first assign a unique prior vector representation to each combination of qualitative inputs. We then use a low-rank linear map to project these priors on a manifold that characterizes the posterior representations. As the posteriors are quantitative, they can be directly used in any standard correlation function such as the Gaussian or Matern. Hence, the optimal map and the corresponding manifold, along with other hyperparameters of the correlation function, can be systematically learned via maximum likelihood estimation. Through a wide range of analytic and real-world examples, we demonstrate the advantages of LMGPs over state-of-the-art methods in terms of accuracy and versatility. In particular, we show that LMGPs can handle variable-length inputs, have an explainable neural network interpretation, and provide insights into how qualitative inputs affect the response or interact with each other. We also employ LMGPs in Bayesian optimization and illustrate that they can discover optimal compound compositions more efficiently than conventional methods that convert compositions to qualitative variables via manual featurization.

42 ENGINEERING↗