Search NASA⌕ Search

SEARCH · Search NASA

Results for “gradient sensing”

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 469 records · Page 26

Balloon-based interferometric techniques

A balloon-borne triple-etalon Fabry-Perot Interferometer, observing the Doppler shifts of absorption lines caused by molecular oxygen and water vapor in the far red/near infrared spectrum of backscattered sunlight, has been used to evaluate a passive spaceborne remote sensing technique for measuring winds in the troposphere and stratosphere. There have been two successful high altitude balloon flights of the prototype UCL instrument from the National Scientific Balloon Facility at Palestine, TE (May 80, Oct. 83). The results from these flights have demonstrated that an interferometer with adequate resolution, stability and sensitivity can be built. The wind data are of comparable quality to those obtained from operational techniques (balloon and rocket sonde, cloud-top drift analysis, and from the gradient wind analysis of satellite radiance measurements). However, the interferometric data can provide a regular global grid, over a height range from 5 to 50 km in regions of clear air. Between the middle troposphere (5 km) and the upper stratosphere (40 to 50 km), an optimized instrument can make wind measurements over the daylit hemisphere with an accuracy of about 3 to 5 m/sec (2 sigma). It is possible to obtain full height profiles between altitudes of 5 and 50 km, with 4 km height resolution, and a spatial resolution of about 200 km, along the orbit track. Below an altitude of about 10 km, Fraunhofer lines of solar origin are possible targets of the Doppler wind analysis. Above an altitude of 50 km, the weakness of the backscattered solar spectrum (decreasing air density) is coupled with the low absorption crosssection of all atmospheric species in the spectral region up to 800 nm (where imaging photon detectors can be used), causing the along-the-track resolution (or error) to increase beyond values useful for operational purposes. Within the region of optimum performance (5 to 50 km), however, the technique is a valuable potential complement to existing wind measuring systems and can provide a low cost addition to powerful active (LIDAR) wind measuring systems now under development.

Rees, David↗

Monitoring Geothermal Features in Yellowstone National Park with ATLAS Multispectral Imagery

The National Park Service (NPS) must produce an Environmental Impact Statement for each proposed development in the vicinity of known geothermal resource areas (KGRAs) in Yellowstone National Park. In addition, the NPS monitors indicator KGRAs for environmental quality and is still in the process of mapping many geothermal areas. The NPS currently maps geothermal features with field survey techniques. High resolution aerial multispectral remote sensing in the visible, NIR, SWIR, and thermal spectral regions could enable YNP geothermal features to be mapped more quickly and in greater detail In response, Yellowstone Ecosystems Studies, in partnership with NASA's Commercial Remote Sensing Program, is conducting a study on the use of Airborne Terrestrial Applications Sensor (ATLAS) multispectral data for monitoring geothermal features in the Upper Geyser Basin. ATLAS data were acquired at 2.5 meter resolution on August 17, 2000. These data were processed into land cover classifications and relative temperature maps. For sufficiently large features, the ATLAS data can map geothermal areas in terms of geyser pools and hot springs, plus multiple categories of geothermal runoff that are apparently indicative of temperature gradients and microbial matting communities. In addition, the ATLAS maps clearly identify geyserite areas. The thermal bands contributed to classification success and to the computation of relative temperature. With masking techniques, one can assess the influence of geothermal features on the Firehole River. Preliminary results appear to confirm ATLAS data utility for mapping and monitoring geothermal features. Future work will include classification refinement and additional validation.

Spruce, Joseph↗

Determining Planetary Boundary Layer Height by Micro-pulse Lidar with Validation by UAV Measurements

Planetary boundary layer height (PBLH) is often used to characterize the structure of the lower atmosphere. Aerosol lidar, a ground-based remote sensing method, provides the vertical distribution of aerosol at a high temporal resolution observation data, from which, the PBL structure and the position of the PBL top can be comprehensively studied. PBLH determination with lidar data depends primarily on the characteristic turbulent motions in the atmosphere and the geophysical location. However, lidar determination of PBLH over densely populated subtropical locations has rarely been discussed; thus, developing retrieval techniques suitable to these areas is necessary. In this study, four PBLH determination methods (Gradient, δ–threshold, Haar wavelet transform, and hybrid image processing) are applied to estimate the PBLH from lidar observations over an urban area in East Asia, and one—the Gradient method—relied on potential temperature measurements from an unmanned aerial vehicle (UAV) flights to validate our results. Our results indicate that a combination of the gradient method and δ-threshold method can provide better results, in terms of diurnal pattern, than using either method individually. Furthermore, the Haar wavelet and the Hybrid image processing can detect the PBL development comparably well, but both methods are dependent on their initial conditions and optimized algorithm settings. In addition, the accompanying UAV observations are conclusively shown to have a high degree of efficacy for validating the lidar data. This research highlights that a combination of PBLH determination methods can better describe the PBLH evolution throughout a day in some cases, while in others less common determination methods are proving useful, and a suite of retrieval methods should still be explored for precisely mapping the PBL in densely populated subtropical areas.

Yueh-Chen Wang↗

Developing flow in S-shaped ducts. 1: Square cross-section duct

Laser-Doppler velocimetry was used to measure the laminar and turbulent flow in an S-duct formed with two 22.5 deg sectors of a bend with ratio of mean radius of curvature to hydraulic diameter of 7.0. The boundary layers at the inlet to the bend were about 25% and 15% of the hydraulic diameter for the laminar and turbulent flows, respectively. Pressure-driven secondary flows develop in the first half of the S-duct and persist into the second half but are largely reversed by the exit plane as a consequence of the change in the sense of curvature. There is, however, a region near the outer wall of the second bend where the redistribution of the streamwise isotachs results in a reinforcement of the secondary flow which was established in the first half of the S-duct. The net redistribution of the streamwise isotachs is comparable to that occurring in unidirectional bends of stronger curvature. The wall pressure distribution was also measured for the turbulent flow and quantifies the expected large variations in the longitudinal pressure gradient distributions which occur at different radial locations.

Taylor, A. M. K. P.↗

Direct Numerical Simulations of Transitional/Turbulent Wakes

The interest in transitional/turbulent wakes spans the spectrum from an intellectual pursuit to understand the complex underlying physics to a critical need in aeronautical engineering and other disciplines to predict component/system performance and reliability. Cylinder wakes have been studied extensively over several decades to gain a better understanding of the basic flow phenomena that are encountered in such flows. Experimental, computational and theoretical means have been employed in this effort. While much has been accomplished there are many important issues that need to be resolved. The physics of the very near wake of the cylinder (less than three diameters downstream) is perhaps the most challenging of them all. This region comprises the two detached shear layers, the recirculation region and wake flow. The interaction amongst these three components is to some extent still a matter of conjecture. Experimental techniques have generated a large percentage of the data that have provided us with the current state of understanding of the subject. More recently computational techniques have been used to simulate cylinder wakes, and the data from such simulations are being used to both refine our understanding of such flows as well as provide new insights. A few large eddy and direct numerical simulations (LES and DNS) of cylinder wakes have appeared in the literature in the recent past. These investigations focus on the low Reynolds number range where the cylinder boundary layer is laminar (sub-critical range). However, from an engineering point of view, there is considerable interest in the situation where the upper and/or lower boundary layer of an airfoil is turbulent, and these turbulent boundary layers separate from the airfoil to contribute to the formation of the wake downstream. In the case of cylinders, this only occurs at relatively large unit Reynolds numbers. However, in the case of airfoils, the boundary layer has the opportunity to transition to turbulence on the airfoil surface at a relatively lower unit Reynolds number because the characteristic length of the airfoil is typically one to two orders of magnitude larger than the trailing edge diameter. This transition to turbulence would occur unless there is a strong favorable pressure gradient that results in the boundary layer remaining laminar or transitional over the surface of the airfoil. This presentation will focus on two direct numerical simulations that have been performed at NASA ARC. The first is of a cylinder wake with laminar separating boundary layers. The second is the wake of a flat plate with a circular trailing edge. The upper and lower plate surface boundary layers are both turbulent and statistically identical. Thus the computed wake is symmetric in a statistical sense. This flow is more representative of airfoil wakes than cylinder wakes. Results from the two simulations including flow visualization and turbulence statistics in the near wake will be presented at the seminar.

wakes↗

Normal mode analysis of macromolecular systems with the mobile block Hessian method

Until recently, normal mode analysis (NMA) was limited to small proteins, not only because the required energy minimization is a computationally exhausting task, but also because NMA requires the expensive diagonalization of a 3N{sub a}×3N{sub a} matrix with N{sub a} the number of atoms. A series of simplified models has been proposed, in particular the Rotation-Translation Blocks (RTB) method by Tama et al. for the simulation of proteins. It makes use of the concept that a peptide chain or protein can be seen as a subsequent set of rigid components, i.e. the peptide units. A peptide chain is thus divided into rigid blocks with six degrees of freedom each. Recently we developed the Mobile Block Hessian (MBH) method, which in a sense has similar features as the RTB method. The main difference is that MBH was developed to deal with partially optimized systems. The position/orientation of each block is optimized while the internal geometry is kept fixed at a plausible - but not necessarily optimized - geometry. This reduces the computational cost of the energy minimization. Applying the standard NMA on a partially optimized structure however results in spurious imaginary frequencies and unwanted coordinate dependence. The MBH avoids these unphysical effects by taking into account energy gradient corrections. Moreover the number of variables is reduced, which facilitates the diagonalization of the Hessian. In the original implementation of MBH, atoms could only be part of one rigid block. The MBH is now extended to the case where atoms can be part of two or more blocks. Two basic linkages can be realized: (1) blocks connected by one link atom, or (2) by two link atoms, where the latter is referred to as the hinge type connection. In this work we present the MBH concept and illustrate its performance with the crambin protein as an example.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Asymmetries in cloud microphysical properties ascribed to sea ice leads via water vapour transport in the central Arctic

To investigate the influence of sea ice openings like leads on wintertime Arctic clouds, the air mass transport is exploited as a heat and humidity feeding mechanism which can modify Arctic cloud properties. Cloud microphysical properties in the central Arctic are analysed as a function of sea ice conditions during the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) expedition in 2019–2020. The Cloudnet classification algorithm is used to characterize the clouds based on remote sensing observations and the atmospheric thermodynamic state from the observatory on board the research vessel (RV) Polarstern. To link the sea ice conditions around the observational site with the cloud observations, the water vapour transport (WVT) being conveyed towards RV Polarstern has been utilized as a mechanism to associate upwind sea ice conditions with the measured cloud properties. This novel methodology is used to classify the observed clouds as coupled or decoupled to the WVT based on the location of the maximum vertical gradient of WVT height relative to the cloud-driven mixing layer. Only a conical sub-sector of sea ice concentration (SIC) and the lead fraction (LF) centred on the RV Polarstern location and extending up to 50 km in radius and with an azimuth angle governed by the time-dependent wind direction measured at the maximum WVT is related to the observed clouds. We found significant asymmetries for cases when the clouds are coupled or decoupled to the WVT and selected by LF regimes. Liquid water path of low-level clouds is found to increase as a function of LF, while the ice water path does so only for deep precipitating systems. Clouds coupled to WVT are found to generally have a lower cloud base and larger thickness than decoupled clouds. Thermodynamically, for coupled cases the cloud-top temperature is warmer and accompanied by a temperature inversion at the cloud top, whereas the decoupled cases are found to be closely compliant with the moist adiabatic temperature lapse rate. The ice water fraction within the cloud layer has been found to present a noticeable asymmetry when comparing coupled versus decoupled cases. This novel approach of coupling sea ice to cloud properties via the WVT mechanism unfolds a new tool to study Arctic surface–atmosphere processes. With this formulation, long-term observations can be analysed to enforce the statistical significance of the asymmetries. Furthermore, our results serve as an opportunity to better understand the dynamic linkage between clouds and sea ice and to evaluate its representation in numerical climate models for the Arctic system.

54 ENVIRONMENTAL SCIENCES↗

A Printed Microscopic Universal Gradient Interface for Super Stretchable Strain‐Insensitive Bioelectronics

Abstract Stretchable electronics capable of conforming to nonplanar and dynamic human body surfaces are central for creating implantable and on‐skin devices for high‐fidelity monitoring of diverse physiological signals. While various strategies have been developed to produce stretchable devices, the signals collected from such devices are often highly sensitive to local strain, resulting in inevitable convolution with surface strain‐induced motion artifacts that are difficult to distinguish from intrinsic physiological signals. Here all‐printed super stretchable strain‐insensitive bioelectronics using a unique universal gradient interface (UGI) are reported to bridge the gap between soft biomaterials and stiff electronic materials. Leveraging a versatile aerosol‐based multi‐materials printing technique that allows precise spatial control over the local stiffnesses with submicron resolution, the UGI enables strain‐insensitive electronic devices with negligible resistivity changes under a 180% uniaxial stretch ratio. Various stretchable devices are directly printed on the UGI for on‐skin health monitoring with high signal quality and near‐perfect immunity to motion artifacts, including semiconductor‐based photodetectors for sensing blood oxygen saturation levels and metal‐based temperature sensors. The concept in this work will significantly simplify the fabrication and accelerate the development of a broad range of wearable and implantable bioelectronics for real‐time health monitoring and personalized therapeutics.

Song, Kaidong [Department of Aerospace and Mechani↗

Neural units with time-dependent functionality

We show that the time-resolved dynamics of an underdamped harmonic oscillator can be used to do multifunctional computation, performing distinct computations at distinct times within a single dynamical trajectory. We consider the amplitude of an oscillator whose inputs influence its frequency. The activity of the oscillator at fixed times is a nonmonotonic function of its inputs, so it can solve problems such as XOR that are not linearly separable. The activity of the oscillator at fixed input is a nonmonotonic function of time, so it is multifunctional in a temporal sense, and able to carry out distinct nonlinear computations at distinct times within the same dynamical trajectory. We show that a single oscillator, observed at different times, can act as all of the elementary logic gates and perform binary addition, the latter usually implemented in hardware using five logic gates. We show that a set of n oscillators, observed at different times, can perform an arbitrary number of analog-to-n-bit digital conversions. We also show that oscillators can be trained by gradient descent to perform distinct classification tasks at distinct times. Computing with time-dependent functionality can be done in or out of equilibrium, and suggests a way of reducing the number of parameters or devices required to do nonlinear computations.

97 MATHEMATICS AND COMPUTING↗

A New Research Approach for Observing and Characterizing Land–Atmosphere Feedback

Forecast errors with respect to wind, temperature, moisture, clouds, and precipitation largely correspond to the limited capability of current earth system models to capture and simulate and-atmosphere feedback. To facilitate its realistic simulation in next generation models, an improved process understanding of the related complex interactions is essential. To this end, accurate 3D observations of key variables in the land-atmosphere (L-A) system with high vertical and temporal resolution from the surface to the free troposphere are indispensable. Recently, we developed a synergy of innovative ground-based, scanning active remote sensing systems for 2D to 3D measurements of wind, temperature, and water vapor from the surface to the lower troposphere that is able to provide comprehensive data sets for characterizing L-A feedback independently of any model input. Several new applications are introduced such as the mapping of surface momentum, sensible heat, and latent heat fluxes in heterogeneous terrain, the testing of Monin-Obukhov similarity theory and turbulence parameterizations, the direct measurement of entrainment fluxes, and the development of new flux-gradient relationships. An experimental design taking advantage of the sensors’ synergy and advanced capabilities was realized for the first time during the Land Atmosphere Feedback Experiment (LAFE), conducted at the Atmospheric Radiation Measurement Program Southern Great Plains site in August 2017. The scientific goals and the strategy of achieving them with the LAFE data set are introduced. We envision the initiation of innovative L-A feedback studies in different climate regions to improve weather forecast, climate, and earth system models worldwide.

54 ENVIRONMENTAL SCIENCES↗

NetFlow Dynamics

NetFlow Dynamics is a web-accessible analysis environment for simulating dynamic flows of materials on model networks. Performing a simulation requires both the NetFlow Dynamics application and a network model which is a description of the structure of the nodes and edges of a network including the flow capacity of each edge and the storage capacity of each node, and the sources and sinks of the material flowing on the network. NetFlow Dynamics consists of databases for storing network models, algorithms to calculate flows on networks, and a GIS-based graphical interface for performing simulations and viewing simulation results. Simulated flows are dynamic in the sense that flows on each edge of the network and inventories at each node change with time and can be out of equilibrium with boundary conditions. Any number of network models could be simulated using Net Flow Dynamics. To date, the models simulated have been models of petroleum infrastructure. The main model has been the National Transportation Fuels Model (NTFM), a network of U.S. oil fields, transmission pipelines, rail lines, refineries, tank farms, and distribution terminals. NetFlow Dynamics supports two different flow algorithms, the Gradient Flow algorithm and the Inventory Control algorithm, that were developed specifically for the NetFlow Dynamics application. The intent is to add additional algorithms in the future as needed. The ability to select from multiple algorithms is desirable because a single algorithm never covers all analysis needs. The current algorithms use a demand-driven capacity-constrained formulation which means that the algorithms strive to use all available capacity and stored inventory to meet desired flows to sinks, subject to the capacity constraints of each network component. The current flow algorithms are best suited for problems in which a material flows on a capacity-constrained network representing a supply chain in which the material supplied can be stored at each node of the network. In the petroleum models, the flowing materials are crude oil and refined products that can be stored at tank farms, refineries, or terminals (i.e. the nodes of the network). Examples of other network models that could be simulated are currency flowing in a financial network, agricultural products moving to market, or natural gas flowing on a pipeline network.

Corbet Jr., Thomas F↗

Less is More: Bigger Data from Compressive Measurements

Compressive sensing approaches are beginning to take hold in (scanning) transmission electron microscopy (S/TEM) [1,2,3]. Compressive sensing is a mathematical theory about acquiring signals in a compressed form (measurements) and the probability of recovering the original signal by solving an inverse problem [4]. The inverse problem is underdetermined (more unknowns than measurements), so it is not obvious that recovery is possible. Compression is achieved by taking inner products of the signal with measurement weight vectors. Both Gaussian random weights and Bernoulli (0,1) random weights form a large class of measurement vectors for which recovery is possible. The measurements can also be designed through an optimization process. The key insight for electron microscopists is that compressive sensing can be used to increase acquisition speed and reduce dose. Building on work initially developed for optical cameras, this new paradigm will allow electron microscopists to solve more problems in the engineering and life sciences. We will be collecting orders of magnitude more data than previously possible. The reason that we will have more data is because we will have increased temporal/spatial/spectral sampling rates, and we will be able ability to interrogate larger classes of samples that were previously too beam sensitive to survive the experiment. For example consider an in-situ experiment that takes 1 minute. With traditional sensing, we might collect 5 images per second for a total of 300 images. With compressive sensing, each of those 300 images can be expanded into 10 more images, making the collection rate 50 images per second, and the decompressed data a total of 3000 images [3]. But, what are the implications, in terms of data, for this new methodology? Acquisition of compressed data will require downstream reconstruction to be useful. The reconstructed data will be much larger than traditional data, we will need space to store the reconstructions during analysis, and the computational demands for analysis will be higher. Moreover, there will be time costs associated with reconstruction. Deep learning [5] is an approach to address these problems. Deep learning is a hierarchical approach to find useful (for a particular task) representations of data. Each layer of the hierarchy is intended to represent higher levels of abstraction. For example, a deep model of faces might have sinusoids, edges and gradients in the first layer; eyes, noses, and mouths in the second layer, and faces in the third layer. There has been significant effort recently in deep learning algorithms for tasks beyond image classification such as compressive reconstruction [6] and image segmentation [7]. A drawback of deep learning, however, is that training the model requires large datasets and dedicated computational resources (to reduce training time to a few days). A second issue is that deep learning is not user-friendly and the meaning behind the results is usually not interpretable. We have shown it is possible to reduce the data set size while maintaining model quality [8] and developed interpretable models for image classification [9], but the demands are still significant. The key to addressing these problems is to NOT reconstruct the data. Instead, we should design computational sensors that give answers to specific problems. A simple version of this idea is compressive classification [10], where the goal is to classify signal type from a small number of compressed measurements. Classification is a much simpler problem than reconstruction, so 1) much fewer measurements will be necessary, and 2) these measurements will probably not be useful for reconstruction. Other simple examples of computational sensing include determining object volume or the number of objects present in the field of view [11].

Stevens, Andrew↗

Molecular shifts in dissolved organic matter along a burn severity continuum for common land cover types in the Pacific Northwest, USA

Increasing wildfire severity is of growing concern in the western United States, with consequences for the production, composition, and mobilization of dissolved organic matter (DOM) from terrestrial to aquatic systems. Our current understanding of wildfire impacted DOM (often termed pyrogenic DOM) composition is largely built from temperature-based studies that can be difficult to extrapolate to field conditions, which are often defined by ‘burn severity’, or the post-wildfire impact observed at a site. Thus, burn severity can encapsulate a broader range of fire and environmental conditions not exclusive to temperature. Biogeochemical studies that describe DOM along burn severity continuums remain limited but are needed to better link DOM composition with field conditions post-fire. Here, in this study, we addressed this need with an experimental open air burn simulation that generated chars from vegetation representative of major land cover types in the western United States. The chars were leached to simulate DOM mobilization potential. The DOM composition was characterized by ultra-high resolution mass spectrometry (HR-MS) and UV/VIS absorbance and fluorescence. Our results indicated that the shifts of DOM production and composition along a burn-severity gradient depends on the land cover type that was burned, with the degree of change dependent on the composition of the starting parent vegetation material. Fluorescence signatures indicated a strong convergence across land cover types to more aromatic DOM with increasing severity, while HR-MS indicated an increase in the production of aromatic nitrogen containing DOM with increasing severity. Results from this study enhance our ability to describe DOM composition in a framework that can be more directly related with field and remote-sensing based metrics.

54 ENVIRONMENTAL SCIENCES↗

An unconditionally-stable central differencing scheme for high Reynolds number flows

The central difference scheme (CDS) is a second order accurate scheme which is free of numerical diffusion (in the second order sense) and is simple to implement: However, for grid Peclet numbers larger than 2, the CDS leads to over- and undershoots and is unstable. The present paper describes a method, called CONDIF, which retains the essential nature of the CDS but eliminates the over- and under-shoots. It leads to unconditionally positive coefficients and, in the limit, approaches the CDS for all values of grid Peclet numbers. The CONDIF modifies the CDS by introducing a controlled amount of numerical diffusion based on the local gradients. In the worst case the scheme yields results similar to those of the hybrid scheme. This paper reports the results obtained from CONDIF for a number of test problems which have been widely used for comparative study of numerical schemes in the published literature. For most of these problems, the CONDIF results are significantly more accurate than the hybrid scheme at high Peclet numbers. In particular, the CONDIF scheme depicts much lower level of numerical diffusion than the hybrid scheme even when the Peclet number is very high and the flow is at large angles to the grid.

Runchal, Akshai K.↗

Satellite Observations of Desert Dust-induced Himalayan Snow Darkening

The optically thick aerosol layer along the southern edge of the Himalaya has been subject of several recent investigations relating to its radiative impacts on the South Asian summer monsoon and regional climate forcing. Prior to the onset of summer monsoon, mineral dust from southwest Asian deserts is transported over the Himalayan foothills on an annual basis. Episodic dust plumes are also advected over the Himalaya, visible as dust-laden snow surface in satellite imagery, particularly in western Himalaya. We examined spectral surface reflectance retrieved from spaceborne MODIS observations that show characteristic reduction in the visible wavelengths (0.47 nm) over western Himalaya, associated with dust-induced solar absorption. Case studies as well as seasonal variations of reflectance indicate a significant gradient across the visible (0.47 nm) to near-infrared (0.86 nm) spectrum (VIS-NIR), during premonsoon period. Enhanced absorption at shorter visible wavelengths and the resulting VIS-NIR gradient is consistent with model calculations of snow reflectance with dust impurity. While the role of black carbon in snow cannot be ruled out, our satellite-based analysis suggests the observed spectral reflectance gradient dominated by dust-induced solar absorption during premonsoon season. From an observational viewpoint, this study underscores the importance of mineral dust deposition toward darkening of the western Himalayan snow cover, with potential implications to accelerated seasonal snowmelt and regional snow albedo feedbacks.

Remote Sensing↗

On-board Radiation Sensor and Low-Altitude Contour Mapping of Radiation Fields using UAS Swarm - 18239

This paper addresses the design of light weight radiation sensors for the small-scale unmanned aerial system (UAS), and its implementation for low-altitude radiation source localization and contour mapping. The compact, high-resolution gamma-ray CZT sensors were integrated to UAS platforms as plug-and-play components using Robot Operation System (ROS). In this work, the UAS swarm consists of three platforms positioned in circular formation. The onboard data analysis provides time and position stamped intensities of gamma-ray peaks for each UAS-based sensor that are used as the input data for the swarm flight control algorithm. Each UAS measures radiation simultaneously for estimating gradient of radiation to compute a bulk heading angle of the swarm for contour mapping, which can provide relatively precise boundary of the safe area for potential human exploration. The UAS radiation sensors allow measurements from just few meters up to hundreds of meters thus filling the gap between land- and helicopter- based platforms. The swarm of UAS has advantages over a single agent based approach in detecting radiative sources and effectively mapping the area. The proposed method can locate sources of radiation as well as mapping the contaminated area for enhancing situation awareness capabilities for first responders. Computer simulation is presented for UAS swarm based source localization and contour mapping in the simulated radiation fields created by the MCNP code. Selected key gradient estimation and contour mapping algorithms are validated experimentally with small scale multicopter UAS in the indoor flight test bed. (authors)

42 ENGINEERING↗

Push it to the limit: comparing periodic and local approaches to density functional theory for intermolecular interactions

With the aim of systematically comparing two popular approaches to density functional theory–all-electron calculations with local basis sets, and periodic calculations employing plane wave basis sets and norm-conserving pseudopotentials–we have computed complete-basis binding energies across the S22 set of intermolecular interactions, a dataset consisting of noncovalent interactions of small- and medium-sized molecules containing first- and second-row atoms, using the Troullier-Martins norm-conserving pseudopotentials with SPW92, a local spin-density approximation; and PBE, a generalised gradient approximation. Here, we have found that it is challenging to reach the basis set limit with these periodic calculations; for the methods and systems examined, a minimum vacuum distance of 30 Å between a system and its nearest images is necessary–unless some form of dipole correction is employed–as is a kinetic energy cutoff of at least 80 Ry. The trends in convergence with respect to vacuum size and kinetic energy cutoff are largely independent of the level of density functional approximation employed. A sense of the impact of each hyperparameter on basis set error provides a foundation for ensuring quality calculations in future studies and allows us to quantify the basis set errors incurred in existing studies on similar systems.

74 ATOMIC AND MOLECULAR PHYSICS↗

Lowering the computational barrier: Partially Bayesian neural networks for transparency in medical imaging AI

Deep Neural Networks (DNNs) can provide clinicians with fast and accurate predictions that are highly valuable for high-stakes medical decision-making, such as in brain tumor segmentation and treatment planning. However, these models largely lack transparency about the uncertainty in their predictions, potentially giving clinicians a false sense of reliability that may lead to grave consequences in patient care. Growing calls for Transparent and Responsible AI have promoted Uncertainty Quantification (UQ) to capture and communicate uncertainty in a systematic and principled manner. However, traditional Bayesian UQ methods remain prohibitively costly for large, million-dimensional tumor segmentation DNNs such as the U-Net. In this work, we discuss a computationally-efficient UQ approach via the partially Bayesian neural networks (pBNN). In pBNN, only a single layer, strategically selected based on gradient-based sensitivity analysis, is targeted for Bayesian inference. We illustrate the effectiveness of pBNN in capturing the full uncertainty for a 7.8-million parameter U-Net. We also demonstrate how practitioners and model developers can use the pBNN's predictions to better understand the model's capabilities and behavior.

97 MATHEMATICS AND COMPUTING↗