137091 research outputs found
Sort by
Signal and Symbol Representations in Sonification: Structuring Data-to-Display Transformations
No full textPresented at the 30th International Conference on Auditory Display (ICAD 2025)Multimodal display systems offer great potential for perceiving and understanding the increasing complexity and volume of data in our contemporary world, enabling better accessibility and interaction. However, increasing data complexity generally results in further design challenges. To address these challenges, we present a structuring method based on signals representing data structures, and symbols representing data contextual qualities. While current research mainly focuses on mapping strategies and common idioms across modalities, our method integrates mappings as part of a broader range of data transformation processes. Through this approach, we structure data transformations to facilitate multiple modalities integration, thus evidencing designers’ decisions throughout a cohesive data-to-display workflow. To evaluate our high-order structure, we developed a sonification of ocean data with complementary visualizations and conducted a formative workshop with two design practitioners (n=2) to gather feedback on its application. Initial results suggest that our structure helps organize the design process, supporting practitioners intent
Strain Localization and Electrical Resistance Evolution in Ag Flake-Based Nanocomposite Inks under Monotonic and Cyclic Stretching
No full textFlexible hybrid electronics (FHE) devices are an emerging class of electronics devices that are flexible and conformal to non-planar surfaces. Interconnect materials for FHE devices known as conductive inks consist of submicron Ag flakes embedded in polymer binder materials and screen printed on polymer substrates. For these conductors, strain localization plays a dominant role in the electrical resistance increase with uniaxial strain. The current work seeks to investigate the origins of ink strain localization, the important material and structural dimensional factors affecting strain localization, and the relationship between strain localization mechanisms and resistance increase with both monotonic and cyclic uniaxial strain. A secondary research objective is the investigation of key empirical parameters for characterizing ink resistance evolution with cyclic strain in order to model ink resistance evolution with cycling.
The primary method for investigating strain localization was uniaxial stretch testing with synchronous resistance measurement. In particular, experiments with in situ scanning electron microscope (SEM) imaging were used to simultaneously monitor the evolution of strain localization (primarily cracking) and electrical resistance with monotonic or cyclic stretching. Different ink and substrate materials were tested using the in situ SEM technique to observe differences in cracking mechanisms. The in situ images were analyzed using digital image correlation (DIC) to generate strain maps, leading to the numerical modeling of resistance based on crack statistics. Experiments with in situ confocal microscope (CM) imaging were used to help understand the effects of ink trace line width and thickness. Finite element modeling was used to investigate the origins of strain localization, which were hypothesized to be surface roughness and local variations in Ag flake concentration. For the second research objective, a series of cyclic experiments helped identify the key empirical parameters, leading to a model for predicting resistance assuming a constant rate of resistance increase with cycling.Ph.D.Mechanical Engineerin
Explaining Individual Differences in the Rate of Focusing Attention: Dissociating Attention Control from Drift Rates
No full textThe top-down control of attention and the speed of information processing are two competing explanations of individual differences in higher-order cognitive abilities, such as working memory capacity and fluid intelligence. However, the relationship between attention control and processing speed remains unclear. While some studies show that the two constructs are related but distinct (e.g., Burgoyne et al., 2023), others contend that the two constructs are virtually indistinguishable (Löffler et al., 2024). The present study adds to this discussion by comparing attention control and processing speed as measured by drift rates from the Ratcliff (1978) drift-diffusion model for predicting working memory capacity, fluid intelligence, and selective attention dynamics in two flanker tasks. Results indicate that individual differences in attention control and drift rates are statistically distinct. Additionally, only individual differences in attention control uniquely predicts individual differences in working memory capacity and fluid intelligence. Finally, attention control and drift rates are sensitive to different processing stages underlying flanker performance, with attention control being more strongly related to the rate of selectively focusing attention and drift rate being more strongly related to perceptual decision making. These results affirm that attention control tasks measure processes related to the focusing of attention that are dissociable from individual differences in differences in drift rate. They also indicate that processes related to selective attention are more broadly related to individual differences in cognitive abilities than individual differences in perceptual decision making processes.Ph.D.Psycholog
Molecular Docking with Trivialized Momentum and Riemannian Diffusion Models: RDM-Lite Dock
No full textStructure-based drug design increasingly relies on computational docking to select vast virtual libraries prior to experimental assays, accelerating the early stages of lead discovery. Conventional docking engines simplify the naturally curved spaces of molecular rotations and torsions into simpler Euclidean approximations, leading to issues like singularities, metric distortions, and ineffective sampling that reduce both accuracy and speed. In this thesis, I present RDM-Lite Dock, a “Riemannian-Lite” retrofit to the state-of-the-art generative docking framework DiffDock. My approach replaces the standard quaternion-based noise injection with Brownian motion directly on the special orthogonal group SO(3), and leverages an analytic, truncated heat-kernel score for rotations. I tested and confirmed the Rotate-Only RDM sampler using the HIV-1 protease–indinavir complex (PDB 1HVR). The resulting initial concept achieves an average RMSD of 2.40 angstroms, a DockQ score of 0.22, and an inference time of 3.2 s with only 28 reverse diffusion steps. These findings confirm that being aware of the manifold during rotation sampling can show RMSD improvement at the cost of DockQ; hence showing a proof-of-concept rather than production-ready model of Riemannian Diffusion in Molecular Docking.UndergraduateComputer Scienc
Improving the Efficiency and Accuracy of Landing Site Error Calculations for Re-Entry Vehicles
No full textSpace missions that involve atmospheric re-entry require thorough landing site analysis for mission safety and success. Monte Carlo simulations are frequently utilized to analyze the uncertainty in the landing site caused by atmospheric conditions that may fluctuate during re-entry, but require time on the order of hours to complete the analysis. Thus, the need arises for a more efficient process that maintains a high computational accuracy. To meet this need, reference tables are generated that allow for a user to look up the landing site uncertainty given a set of re-entry conditions. These tables are generated by transforming a traditional re-entry state vector to a modified state vector that contains the magnitude of the re-entry velocity, three angles, and the vehicle’s ballistic coefficient. Then a Monte Carlo simulation is performed for each feasible set of re-entry conditions. A table is first generated for ballistic, non-ablative re-entry vehicles that have no flight control system. Then, a simple propulsive control system is implemented to observe how the control system impacts the resulting landing uncertainty. To improve the applicability of these tables, an inverse-distance weighting interpolation scheme is employed that utilizes the table entries to determine the landing error for any set of re-entry conditions that a user may provide. Comparing this interpolation to a standard Monte Carlo simulation, it is observed that both methods achieve similar results, but the interpolation method requires less than one second compared to the hour required for a standard Monte Carlo simulation. Thus, the work presented in this thesis represents the first intuitive and computationally inexpensive method to determine landing site error without sacrificing the fidelity of the results.M.S.Aerospace Engineerin
CFD Analysis and Molecular Kinetics in a Supercritical CO2 Circuit Breaker
No full textIn this work, the behavior of supercritical carbon dioxide (scCO2) at high pressures is investigated in regards to its arc quenching capabilities to suppress reignition in high-voltage circuit breakers. scCO2 is a favorable replacement for the current SF6 gas, which produces toxic byproducts. SF6 contributes heavily to the global warming crisis, and it is paramount to find a more environmentally conscious alternative. One possibility is scCO2, which has a comparable dielectric strength to SF6 and shows promise in arc quenching applications; however, the behavior of scCO2 is not as well researched as that of SF6, leading to a critical gap in knowledge needed to implement this change. To bridge this gap, a preliminary investigation of scCO2 was done using a novel Particle-In-Cell (PIC) solver approach as opposed to a more classic, continuum flow approach. This work presents the initial results from the simulation program Charge Plus[1], developed by Electro-Magnetic Applications Inc. (EMA3D®). Charge Plus is a Particle-In-Cell (PIC) solver that, rather than using fluid properties, instead utilizes reaction probabilities to track macroparticles representing the flow of species in the simulation. While a continuum flow solver is a suitable approach to handle the initial quenching of the arc, the PIC method is particularly suited to answering this work’s overarching question of whether restrike (or reignition) will occur due to its leverage of molecular kinetics, as restrike is largly a molecular kinetics problem rather than a thermodynamics problem. Through a series of parametric studies, it was determined that scCO2 shows promise preventing arc restrike, but some numerical instabilities and a lack of validation data prevents this work from reaching a conclusive answer to the objective. However, significant progress has been made in building and testing a working model not only for scCO2 but also for a nozzle configuration.M.S.Aerospace Engineerin
A Survey on the Desirability, Usability, and Trust of Advanced Driver Assistance System (ADAS)
No full textADAS technology is becoming more common in vehicles. The goal of this project is
to understand the general population’s perception of ADAS features performing at
diƯerent levels of control using surveys. All surveys were performed in person in the
Atlanta metro area. The survey found that respondents find features they perceive to
perform less function, more desirable and trustworthy. The biggest limitation
respondents find about ADAS features are relinquishing control and a fear of failure.
The biggest benefits respondents identified of ADAS features are safety and
convenience.M.S.Civil Engineerin
Envisioning Technology-Mediated Futures of Care Work
No full textHealthcare delivery and care work are increasingly being seen as areas where information and communication technologies (ICTs) and artificial intelligence (AI) technologies could play key roles in streamlining existing processes and improving current care workflows. Prominent efforts in this space have taken the form of patient-generated health data and other health tracking approaches, patient empowerment through ICTs, remote care technologies, online health, and caregiver support, all largely taking a predominantly patient-centered approach to design. At the same time, emerging technologies like conversational and generative AI, wearable devices, and smart home technologies have all been described as potentially revolutionary advances in care work. They promise to automate and transform the overall care experience for caregivers and patients, and are receiving increasing attention and investment from both academia and industry. This comes at a time when there is an acute demand for improved care infrastructures around the world in the wake of the COVID-19 pandemic. As a result, there is an urgent need to understand not only how such technologies support and enable care work, but also how they can be integrated meaningfully into culturally situated care workflows. When done well, this could serve to redistribute the burdens of care work onto technologies while maintaining patients' and caregivers' agencies in care and ensuring continued provision of patient-centered care.
In this dissertation, I examine different contexts of care and how technology-mediated workflows are being employed in care provision in urban India. I employ ethnographic and qualitative research methods, and technology probes to gain insights into how various stakeholders in care ecologies are using ICTs in their care practice. My dissertation brings us to a point where we understand how existing care workflows are attempting to leverage technologies towards meeting the aspirational goals of patient-centered care, and how and where they are facing hurdles. I contribute rich descriptions of how various ecological actors leverage technology-supported workflows in their everyday care work at home, in clinical settings, and in public health. This dissertation further contributes empirical insights into how these practices are influenced, informed, and constrained by sociocultural norms and socioeconomic realities, and what it might mean to better approach technology integration into care workflows. Through design recommendations and implications for each of the three contexts of care, my dissertation helps envision how new technology-mediated practices might comprise more equitable, accessible futures of care work.Ph.D.Human – Centered Computin
Assessing Micromobility as a First- and Last-Mile Solution to Improve Transit Accessibility for Low- Income Communities
No full textThis paper is an applied research paper for a master of city and regional planning; the document is structured into seven main chapters, including introduction, methodology, results, discussion, and references.This study examines how micromobility services—specifically shared bicycles and e-scooters—can enhance first- and last-mile (FM/LM) connectivity to public transit and improve access to employment opportunities in the Atlanta metropolitan region. Although micromobility has gained recognition as a flexible multimodal solution, important uncertainties remain regarding how different modes compare in terms of spatial availability, accessibility improvements, and equity outcomes across neighborhoods with varying socioeconomic conditions