137091 research outputs found
Sort by
Crafting Social Robots for Neurodiverse Individuals: Prospects and Challenges
No full textThis study explores the potential of robotic dogs as tools for neurodivergent individuals in therapy, focusing on their socio-emotional well-being. The limitations of traditional animal therapy were considered, as we investigated the design and impact of robotic dogs as another option. Two robot dog models, Aibo and A1, were introduced to neurodivergent students in a university setting, putting them in various workshops to assess their responses, preferences, and difficulties. The study's methodology included co-design workshops where students interacted with the robot dogs. They then created use case storyboards and addressed potential robot failures. Through these interactions, students' perceptions of the robots' physical appearance, responsiveness, and potential use cases in daily life were examined. The findings revealed a preference for robot dogs combining Aibo's dog-like and more approachable appearance with A1's responsive behavior. They didn’t appreciate Aibo’s lack of quick decision making and A1’s robotic and unnatural look. Students wanted robot dogs in diverse roles, from companionship and emotional support to practical tasks like safety monitoring. A significant part of the study involved understanding student responses to robot malfunctions or failures. Students expressed a desire for clear communication from the robots during failures, indicating a need for intuitive and empathetic design in robotic companions.UndergraduateComputer Scienc
Detecting the Undetectable: Using Laser Interference Patterns to Study the Biophysics of Antibiotic Resistance
No full textMany species of bacteria have become resistant to previously clinically effective antibiotics, or even resistant to multiple drugs. Despite the growing urge for research in the field, antibiotic resistance is typically studied on agar plates that do not resemble the interior of the human body. Researchers have recently created a medium that closely models the sputum of cystic fibrosis (CF) patients, in whom the infection is a major cause of lung disease due to the patients’ dulled immune response and the bacteria’s mobility. This mobility is tied to the underlying biophysics, the way that bacteria traverse the medium in which they replicate. However, this synthetic CF sputum media (SCFM2) is so opaque that common methods of detecting bacterial growth cannot be utilized. Thus, an alternative technique appropriate for studying bacterial growth in this highly useful medium needs to be developed. One such method is imaging the motion of bacteria via laser speckle contrast imaging (LSCI). Studying bacteria-antibiotic interactions in a realistic medium is vital for understanding bacterial dynamics. From a drug development viewpoint, it is critical to understand bacterial physical properties and resistance mechanisms, including the release of chemicals, binding mechanics, and the bacteria’s movement, the latter being the focus of this work. Some research has already demonstrated that LSCI is an effective method for studying bacterial kinetics in traditional media. As a low-cost technique with high spatial resolution, LSCI could broaden access to bacterial motion studies in resource-limited labs. Hence, the development of an appropriate laser speckle imaging method for SCFM2 could allow for more accurate and cost-efficient study of bacterial dynamics within a more realistic medium. This study aims to evaluate the effectiveness of LSCI in capturing motility topography of bacterial populations in SCFM2.UndergraduatePhysic
Analyzing and improving electronic structure calculations of catalytic interfaces using density functional theory and machine learning
No full textDensity functional theory (DFT) plays an important role in heterogeneous catalysis by enabling first-principles study of large and periodic systems due to its accuracy and low computational cost. It provides detailed insights into the catalyst activity and selectivity at atomic scale. These calculations require an input for the exchange-correlation (xc) functionals, which accounts for the multi-particle interactions in DFT but its universal form is unknown. The most accurate xc functional typically depends on the type of chemical system, making it challenging to choose a functional for systems that contain interfaces between different phases of matter or where multiple types of chemical bonding are important. Semilocal functionals are typically used to calculate chemisorption energies due to their low cost, but they differ from experimental values by as much as 1 eV, which can lead to quantitatively and qualitatively incorrect conclusions in the analysis of surface reaction systems. In this thesis, we first explore the typical model space of xc functionals: hybrid and generalized gradient approximation (GGA) level functionals in DFT to investigate the convergence of surface properties and electronic gap of rutile titania nanoparticles with particle size. The geometric and electronic finite-size effects in surface energy are deconvoluted and the influence of defects on electronic gap are evaluated. Further, we explore a novel approach for xc functional design using the multipole (MP) descriptor family to describe the local electronic environments in chemical systems. MP descriptors for electron density provide a set of complete, translationally, and 3D-rotationally invariant convolutional descriptors. Utilizing the MP descriptors, we propose a data-driven framework that uses energies from two different levels of theory to predict the gas-phase corrections in heterogeneous catalytic systems. Next, we introduce a new method to construct xc functionals using convolutions of arbitrary kernels with electron density. We derive the variational derivative of these functionals and provide equations for variational derivatives based on MP descriptors from convolutional kernels. A proof-of-concept functional, PBEq which allows a single functional to use different GGAs at different spatial points in a system is implemented. Testing of PBEq functional on small molecules, bulk metals, and surface catalysts suggests that this approach is a promising route to simultaneously optimize multiple properties of interest. Finally, we propose a framework for developing surrogate hybrid functionals using MP descriptors, at the cost of semilocal functionals. This framework highlights the challenges related to satisfaction of physical constraints while linking exact exchange to semilocal functionals. This approach, combined with the derivation of variational derivatives, has the potential to pave the way for new strategies for functional design and integration of self-consistent ML functionals within DFT.Ph.D.Chemical and Biomolecular Engineerin
Characterization of visuomotor encoding in the human premotor eye field (PEF) using intracortical neural recordings
No full textQuadriplegia, a condition characterized by the loss of motor function in all four limbs, significantly impacts the daily lives of those affected. Brain-computer interfaces (BCIs) have emerged as a promising solution to restore function by decoding neural signals. This study investigates the potential for decoding eye movements alongside hand movements, utilizing intracortical neural recordings from the premotor eye field (PEF) and area 55b, a speech related region. A participant in the BrainGate2 clinical trial with an implant in the PEF/55b was instructed to perform a task separating eye and hand movements. Neural activity was recorded using a microelectrode array, and the data was processed to identify modality and direction encoding. Results show that the PEF encodes hand and eye movements modalities distinctly. The PEF region also encodes direction consistently across modalities. These findings indicate that the PEF/55b region may play a role in coordinating hand and eye movements. This work may lead to improved BCI design by considering eye movement alongside hand movements, potentially improving multimodal control for individuals with severe motor impairments.UndergraduateNeuroscienc
Label-free Deep-Ultraviolet Microscopy: Accessible Molecular Imaging from Bench to Point of Care
No full textImaging with ultraviolet (UV) light (wavelengths ranging from ~200-400 nm) enables label-free molecular imaging due to the distinctive absorption and dispersion properties of several physiologically important, endogenous biomolecules in this spectral region. In addition, the shorter wavelength of UV light offers higher spatial resolution than conventional imaging systems that use visible light. Furthermore, advances in UV light sources and detectors have resulted in setups that enable contiguous imaging of live cells over long durations without significant photodamage. This dissertation aims to enhance the capabilities of deep-UV microscopy for accessible imaging of biological samples. Initially, we introduce a simple technique for hyperspectral UV microscopy to extract quantitative absorption information from biological samples without prior knowledge of their optical properties. Following this, we employ multi-spectral deep-UV microscopy to quantify hemoglobin in red blood cells. Subsequently, we leverage recent advances in deep learning to develop an automated pipeline for label-free hematology analysis using single-wavelength UV microscopy images. In conjunction with a compact deep-UV microscope and custom microfluidic devices, this work can enable low-cost, efficient, and label-free hematology analysis within minutes, suitable for clinical, at-home, or low-resource settings. Additionally, we explore the development of a multispectral UV microscope for high-resolution, 3D tomographic imaging of cells. Overall, this dissertation advances UV microscopy through improved instrumentation, analysis, and computational reconstruction, establishing it as an effective and economical label-free imaging tool for research, clinical, and point-of-care applications. We anticipate that the high-resolution molecular and structural information obtained from UV microscopy will further our understanding of fundamental biology and aid in disease diagnosis, monitoring, and treatment planning.Ph.D.Bioengineerin
Advanced Techniques in UV Microscopy: Integrating Deep Learning for Autofocusing, Whole Slide Imaging Optimization, and Spicule Detection in Bone Marrow Aspirations
No full textDeep-ultraviolet (UV) microscopy enables label-free, high-resolution, quantitative molecular imaging and enables unique applications in biomedicine. This is achieved by
leveraging the unique absorption of different biomolecules in the deep UV spectrum (200~400nm). While UV microscopy has historically been limited by phototoxicity concerns, recent advances in UV imaging hardware have allowed for live cell imaging for more than six hours with no observable photodamage. This proposal aims to expand the capabilities of UV microscopy for biomedical applications using CNN based architectures. Initially, we will employ a custom benchtop
multispectral microscope for single-shot autofocusing. Subsequently, we will utilize a lowcost compact UV microscope for automated whole-slide imaging, incorporating singleshot autofocusing as well. Additionally, the compact UV microscope will be employed for automated spicule detection in bone marrow aspirations to enable an untrained clinician to
use the device.M.S.Biomedical Engineerin
Sensory Integration in Balance Perception in Post-Stroke Individuals
No full textStroke leads to somatosensory perceptual deficits that impair whole-body motion (WBM) as well as deficits in overall balance. Literature suggests that impaired perception during WBM may result in decreased overall balance ability, but the underlying neural mechanism between the two is unknown. Investigating the role of cortical sensory processing may help inform a connection between impaired balance and impaired WBM perception after stroke. We hypothesized that reception of sensory information in somatosensory cortex is impaired after stroke, which disrupts balance and WBM perception. We elicited somatosensory evoked potentials (SEPs) to measure cortical sensory processing and compared between controls and participants with stroke in both sitting and standing conditions. We also conducted tests of perceptual ability and overall balance ability and compared performance in those tasks to latency and amplitude of SEPs. Results showed that latency is delayed on the paretic side of participants with stroke, but not between the left and right side of controls. SEP amplitude was larger on the paretic side than the non-paretic side in participants with stroke but showed no side differences for control. Amplitude of SEPs were larger in sitting than in standing across all conditions and sides. However, delayed and weakened cortical sensory processing did not correlate with impaired overall balance or perceptual abilities, indicating that cortical sensory processing is most likely not the link between the two after stroke.UndergraduateNeuroscienc
An Operational and Transparent VISSIMTM Calibration Method for Transportation Professionals in Georgia
No full textVISSIM (hereinafter VISSIM) is one of the premiere microscopic traffic simulation software packages available on the market, and has been used on projects worldwide. PTV, the company which produces, maintains, and licenses VISSIM provides US users with default car- following model parameter values, desired speed distributions, vehicle compositions, etc. Though these may represent a good starting point for any simulation effort within the country, the variety and plurality in the nature, characteristics, and behavior of traffic across different regions mean that for any simulation to be accurate, default parameter values may need to be calibrated to field conditions.
This thesis seeks to contribute to the growing body of knowledge in both the research and practice realms of VISSIM calibration. This thesis proposes a novel approach to the complex calibration problem that tries to bridge the gap between standard procedures in transportation practice and calibration efforts conducted at academic and research institutions. By focusing on what the state of the practice (i.e., calibration procedures used by transportation professionals) and the state of the art (i.e., calibration procedures devised as part of research efforts) are lacking, the proposed calibration method provides a procedure that is operational, transparent, reasonable in data requirements, procedural, and flexible.
After a thorough literature review in Chapter 1, Chapter 2 focuses on proper model building principles, which are essential preparatory steps modelers must conduct before calibrating any model. Chapter 3 then presents a general calibration method, along with a discussion of the guiding principles. Chapters 4 and 5 respectively explore two applications of the general calibration method: interrupted flow facilities (such as arterials and other signal-controlled roadways, in Chapter 4), and uninterrupted flow facilities (such as freeways, in Chapter 5). Finally, Chapter 6 presents some conclusions and future research directions.M.S.Civil Engineerin