21978 research outputs found
Sort by
Long Short-Term Memory Networks to Improve Aerodynamic Coefficient Estimation for Aerocapture
Get PDFAerocapture is a method for orbital insertion from a hyperbolic trajectory being considered for NASA’s proposed 2030’s flagship mission to Uranus. By traveling through the planet's atmosphere to generate drag, aerocapture greatly reduces the fuel needed when firing retrograde thrusters, allowing for larger payloads or a less powerful launch vehicle. Despite these theoretical benefits, aerocapture has never flown on any planetary missions due to thin margins of error and in-situ corrections necessary to properly execute the maneuver. Critical to the guidance and control algorithms are the aerodynamic coefficients. We propose using a neural network to learn the nonlinear relationship between the raw sensor data and these aerodynamic coefficients. Specifically, we explore how network architectures designed for time dependent data, like Long-Short Term Memory (LSTM) neural networks, can produce aerodynamic coefficient estimates akin to that of a computational fluid dynamics (CFD) based lookup table, while providing more robust coefficient estimation when large environmental perturbations are experienced. Improving force and moment coefficient estimation would improve aerocapture by providing more accurate aerodynamic coefficients for use in guidance and control algorithms. This work considers multiple sensed data sources and aerodynamic coefficient data along with LSTM network architectures for model training to maximize an aerocapture maneuver’s success rate when tested in a Monte Carlo simulation. Along with this, sensitivity analyses were conducted on model hyperparameters to account for relationship complexity. Results were compared against traditional aerodynamic coefficient lookup tables within the Fully Numeric Predictor-corrector Aerocapture Guidance (FNPAG) algorithm to draw conclusions for the model’s performance.</p
Gaussian Process Regression for Local sUAS Wind Prediction
Get PDFThe focus of this work is the implementation of Gaussian process regression for continuity of measurement for small uncrewed aircraft systems (sUAS). This method supplies a continuous estimate with uncertainty of the three-dimensional wind vector along a predicted flight trajectory. The Gaussian process is implemented with GPyTorch, which is derived from the popular machine learning Python library, PyTorch.
Wind observations come from the TORUS-LItE dataset collected with the RAAVEN platform. An onboard five-hole probe measures the air-relative velocity vector, from which the wind can be calculated using the GPS velocity. This absolute wind from the multi-hole probe is treated as the ground truth during validation.
Several Gaussian processes are compared with different input and output parameterizations. The inputs include normalized latitude, longitude, altitude, time, autopilot eastward wind, and autopilot northward wind. The kernel used is the squared-exponential radial basis function with separate length-scales for each input dimension. The outputs are the three-dimensional wind vector. A batch independent approach is compared to a multi-task Gaussian process that uses correlated outputs with a learned inter-task (output) covariance matrix. Hyperparameters are learned by minimizing the negative marginal log-likelihood for 50 training epochs using the Adam optimizer.
Each flight is subdivided into a series of rolling windows, and predictions over a 30 second horizon are generated with the previous 30 seconds of data used for inference. Predictive uncertainty is derived from the posterior multivariate normal distribution. The root-mean-square error is calculated for 1, 5, 10, and 30 second forecasts.
A second sUAS flight track is introduced to test whether temporally and spatially coincident observations from a nearby sUAS can improve local predictions. The supplemental RAAVEN aircraft is separated from the primary RAAVEN by approximately 100 m in altitude, and the supplementary RAAVEN's multi-hole probe wind components are added to the Gaussian process input vector.
Results show that predictions up to 5 seconds have low error throughout all the flights. Adding a second sUAS shows promising results for spatially coincident observations but provides little value from temporally coincident observations. The uncertainty in the predictions is quantified and compared over various short term time horizons.</p
Multiscale Infectious Disease Dynamics: Linking Epidemiology and Testing for Outbreak Neutralization
Get PDFInfectious disease dynamics are shaped by interactions across multiple scales. Within a host, pathogens interact with cellular and immune processes, driving fluctuations in pathogen concentrations and immune responses. Between hosts, infectious contacts facilitate pathogen transmission, driving fluctuations in case counts and population immunity levels. However, traditional mathematical models often focus on only one of these scales. This dissertation addresses this gap by leveraging multi-scale modeling to examine how within-host dynamics influence population-level transmission. By integrating models of within-host viral and immune dynamics with between-host epidemiological transmission models, this work provides insights into infectious disease spread and intervention strategies.
First, a multi-scale model is developed to estimate testing effectiveness, the reduction in transmission due to testing and subsequent isolation, using a probabilistic framework which incorporates viral kinetics, test attributes, and testing behaviors. This model provides a general framework for comparing testing strategies for any virus. Second, these results are incorporated into a compartmental modeling framework to analyze the effectiveness of vaccinate-or-test policies for COVID-19. Lastly, a framework is developed to evaluate our ability to learn about correlates of protection by linking immunological marker concentrations with observed infection events in test-negative design studies. This research highlights the importance of incorporating both within-host and between-host processes to understand infectious disease dynamics and evaluate intervention strategies.</p
Seeing Through Label-Tinted Classes: Labels Enhance Perceptual Warping During Novel Category Learning
Get PDFLearning to classify objects has been shown to have a measurable influence on our perceptual experience, such that objects belonging to the same category begin to appear more similar to one another and objects from different categories begin to appear more distinct from one another. This pattern of perceptual warping is known as categorical perception. Linguistic labels have been shown to facilitate category learning, such that novel category learning occurs at a faster rate when labels representing category membership accompany the stimuli to be classified. The primary aim of the work presented in this thesis is to investigate the relationship between this label-augmented category learning effect and the perceptual warping effect. I hypothesized that the accelerated rate of learning new categories in the presence of labels would be accompanied by a corresponding enhancement of perceptual warping effects over and above those that occur in category learning without labels. I test this hypothesis in a series of three studies comprising a total of six experiments. In the first study, I identify a category learning task and set of stimuli that are conducive to producing a reliable label-augmented category learning effect. In addition to replicating the labeling advantage, the results of this study suggest that label-augmented category learning is dependent on certain aspects of the learning context. In the second study, I investigate whether perceptual warping is enhanced during the course of label-augmented category learning by incorporating a same-different task interleaved around blocks of the category learning task. The results of this study suggest that label-augmented category learning coincides with an enhancement of perceptual warping, but are ultimately inconclusive. In a third and final study, EEG data is used to further investigate the time course of the emergence of perceptual warping effects. This study provides evidence that more conclusively supports the hypothesis that perceptual warping effects are enhanced by labels, converging with the pattern of results observed in study two. These results demonstrate that the label-augmented category learning effect coincides with an enhancement of perceptual warping, suggesting that the formation of labeled category representations entails further warping of perceptual space.</p
Allies or Pacific Rivals: Discursive Shifts in How American Foreign Correspondents Covered the U.S.-Japan Relationship, 1945-1975
Get PDFThis thesis considers an often-overlooked historical source in foreign correspondents to gain a better understanding of American perceptions of Japan from the immediate postwar period to 1975. In this period, foreign correspondents grappled with Japan’s future in relation to the United States and considered whether Japan could be an American ally or if a pacific rivalry between the two nations would return. By looking at the Tokyo Bureau Chief of some of the United States’ most popular national newspapers between 1945 and 1947, this thesis makes an argument in two parts. First, foreign correspondents working for the nation’s most prominent newspapers understood the U.S.- Japan relationship primarily through the lens of economic and security issues. Second, through observing the period between the end of World War Two and the middle of the 1970s, foreign correspondents initially affirmed a postwar narrative that suggested Japan’s position as a strong Cold War ally, but in a subordinate position to the United States. However, as American Cold War foreign policy was increasingly scrutinized within and outside of Japan, the climate of U.S. foreign relations forced foreign correspondents to reconsider Japan’s position in relation to the United States. This thesis hopes to add diversity to our understanding of how Americans viewed Japan’s evolving geopolitical position in the postwar period and gain a better knowledge of the processes information went through when transferring across the Pacific Ocean during the golden age of foreign correspondents.
Keywords: foreign correspondents, discursive history, U.S.-Japan relations, Cold War journalism, Allied Occupation of Japan, World War Two legacies</p
What’s in a Cheese?: The Sociosemantic Distinction of Parmigiano and Parmesan through Historical Romance Diachrony and Contemporary American Marketplace Semiotics
Get PDFParmigiano Reggiano DOP is novel in that it is the only notable product of protected designation of origin (PDOP) to have both its origin-language name (Parmigiano) and a phono-semantic match (Parmesan) protected by EU law. In the US, this legal protection is reserved for Parmigiano but not Parmesan.
This study argues, fundamentally, that a sociocultural split spurred a phonotactic split which spurred a semantic split which finally spurs a semiotic split. The aim of this study is to bridge varying fields of general linguistics for a greater understanding of sociolinguistic language use and change. This study relies on Jurafsky et al.’s work on natural and traditional authenticity and Silverstein’s work on indexical sign-object relationships.</p
Good Dogs
Get PDF Good Dogs is a Creative Nonfiction novella that moves between a father’s woodworking
shop and the hidden world of the federal bomb squad. In one story, a son learns the patient craft
of shaping wood under the silence of heirloom hand tools and a father’s steady discipline. The
workshop becomes both a classroom and a crucible, a place where the laws of precision, control,
and patience are passed from master to apprentice—even if in conflict, the dedication to craft and
family persists above all else. In the other, the same son—years later—serves as an Explosive
Ordnance Disposal Operator on the private security detail for Pope Francis’s 2015 U.S. visit,
charged with tasks the public never sees, unfolding in silence and sustained by the trust between
operators and the responsibility for strangers who will never even know they were there. The two
narratives collide, showing how apprenticeship and survival each demand precision, silence, and
the ability to exist inside systems of control.
At its core, Good Dogs is about the compromises that shape us in family and the
institutions we serve. It asks what happens when duty collides with conscience, and whether the
choices we make within can truly constitute agency.
</p
Using the General Lake Model and Shared Socio-Economic Pathways to Understand Climate Change Impacts on Lake Evaporation
Get PDFIt is critical to understand how climate change will impact our world. Lake evaporation will be
impacted by climate change, having large implications on water supply and demand for drinking,
irrigation, and recreational water uses. Understanding how lake evaporation changes under
different climate change scenarios is critical for proper water management and policy. Using the
General Lake Model, I simulated lake evaporation at Standley Lake reservoir in Westminster,
Colorado, under four different climate change scenarios using the SSP1-2.6, SSP2-4.5, SSP3-7.0,
and SSP5-8.5 scenarios for June, July, and August in 2020. In my study, I looked specifically at
changes to air temperature, relative humidity, and wind speed, which are meteorological variables
known to have large impacts on evaporation at the lake surface. These variables were tested
individually and then coupled together. I found that with only changes to air temperature, the total
evaporation decreased due to the decreasing vapor pressure gradient over the study period; changes
to just the relative humidity showed an increase in evaporation between scenarios; and when only
wind speed was changed, lake evaporation increased or decreased when wind speed increased or
decreased, respectively. When the study variables were coupled together, the total evaporation
decreased over the first three scenarios and increased for the last scenario. This points to the change
in wind speed being the largest driver of lake evaporation at Standley Lake over the study period
and highlights the significance of considering more than just air temperature in future water policy
plans and discussions. This study also shows the value in isolating meteorological variables to see
the extent of their impact on lake evaporation.</p
Object & Subject: A Look at the Cinematic Puppet
Get PDFThis thesis examines the nature and use of puppets in cinema, specifically in Jim Henson’s children’s fantasy films The Dark Crystal (Jim Henson and Frank Oz, 1982) and Labyrinth (Jim Henson, 1986). The phenomenon of the puppet is considered from the perspectives of both object and subject. The function of the puppet as object is literal, it being a physical object explicitly manufactured for performance. The concept of subject explores the puppet as character but also considers the roles of both performer and audience in endowing the object with life.
Chapter One defines the concept of puppet and explains how Jim Henson’s Muppets build and expand on the history of puppetry by being designed specifically for the screen. A discussion of duality and reflexivity in The Muppet Movies introduces the dual nature of the puppet as object and subject. Chapter Two focuses on puppet as object through The Dark Crystal, examining scale and detail in the material construction of its puppets. The materiality of the puppets is informed by the world building of this fantasy film, the entire universe is designed from scratch including its characters. Chapter Three focuses on puppet as subject through The Labyrinth, exploring the blend of human and material characters. The idea of puppet as subject is linked to the coming-of-age themes of the film, as the characters and settings can be read as projections of the teenaged protagonist.</p
Foundations of High-Performance Computing Micro-credential Checklist - Kim Wilkins
Get PDFThis micro-credentialed course provides a foundation for addressing computing-, memory-, or storage-intensive research problems using high-performance computing (HPC). Participants who complete the course will be able to navigate the Linux command line, apply data transfer protocols, find and use software on HPC, and use a scheduler to run batch and interactive jobs. Skills acquired in the course can greatly accelerate problem solving in the computational realm. </p
Data sources:
Access Repository Dashboard
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇