35180 research outputs found
Sort by
AN END-TO-END APPROACH FOR SUPPORTING NAVIGATION OF AUTONOMOUS VEHICLES IN RURAL ENVIRONMENTS
Get PDFThe past decade has shown a rising interest in autonomous driving (AD), and the future applications for it continue to expand into new territories. While there has been some work done for rural environments, the majority of the current work has been focused on urban environments. Rural environments present unique challenges to the algorithms onboard an autonomous vehicle (AV) which hinder them from being deployed sooner in these places. All AVs must be equipped to handle a myriad of environmental variables such as lighting, road conditions, emergency traffic maneuvers, and addressing the presence of natural and man-made objects. Urban settings have luxuries that rural settings are not guaranteed to have. They typically have clean, lane-marked, asphalt roads with plenty of traffic lights and traffic street signs to determine what to actions to take. Many cities also have fast and reliable cellular service, enabling “vehicle-to-everything” (V2X) communication between the vehicle and the traffic infrastructure, cloud databases, and other vehicles with V2X technology on the road. On the other hand, in rural conditions, the road may be unpaved or in poor conditions. In scenarios like agricultural settings, there may hardly be a road at all, with several agricultural vehicles being required to drive on dirt and grass. Rural conditions also do not have as much formal traffic infrastructure in order to help guide the flow of traffic. In addition, depending on the location, the AV’s reliability on V2X communication may be hindered by the lack of cellular infrastructure. In order to deploy autonomous vehicles into any terrain, they must be fully equipped edge computing devices that do not require cloud technology. This means that all computation - all the sensing, perception, motion planning, communication, and other algorithms - must be performed in real time onboard the vehicle to ensure safe and reliable driving. This research proposes an end-to-end approach that will enable further development in deploying AVs in rural environments. Current rural AD research have proposed algorithms to utilize segmentation machine learning models in order to navigate rural areas. The issue is that these algorithms require a framework in order to be fully deployed into a real driving scenario. This research aims to address that issue. A prototype case study is shown that demonstrates how segmentation machine learning models can determine where an AV should drive. In this prototype, we utilize a Jetson Nano edge device and optimize the performance of the model with two optimization compilers. The optimized model then performs inference on images of various rural scenarios from a front-facing dash camera view. The prototype is then expanded upon to be used in a full-scale autonomous driving platform known as Autoware®. This distributed computing software package breaks down each component of AVs into separate computation nodes. The nodes for the vehicle’s perception, planning, control, etc. compute information synchronously. In this research, we deploy a YOLOv11 segmentation model into the Autoware perception module. The model then performs real-time inference on a vehicle in a driving simulator called AWSIM. As the vehicle travels in the simulator, the model is able to keep up with the vehicle’s surroundings. This case study is then concluded with how autonomous vehicles are able to use the information gained from the segmentation models to ensure safe driving in rural environments
THE EVALUATION OF ZEOLITE (CLINOPTILOLITE) INCLUSION ON ENTERIC METHANE IN AN ARTIFICIAL RUMEN SIMULATION TECHNIQUE (RUSITEC)
Get PDFZeolite Clinoptilolite (ZC), a naturally occurring mineral, has shown the potential to reduce methane emissions in dairy cattle. With increasing pressure on the agricultural industry to implement methane mitigation strategies, ZC offers a possible solution by targeting ruminal microflora responsible for methane production. This study evaluated ZC as an enteric methane reducer in an artificial closed fermentation system. We hypothesized that ZC could reduce methane while maintaining other production-associated parameters.We utilized 18 fermenters with 9 fermenters per treatment and three runs. Two diets were formulated with adjusted levels of ZC (0% ZC, and 5% DMI ZC). Rumen fluid used as inoculum was collected from harvested cows. The system was consistently supplied with artificial saliva (dilution rate of 2.9%/hr), periodically agitated, fed daily, and kept O2-free to mimic rumination. Effluent produced by the fermenters was measured and recorded daily. Each of the three runs lasted 12 days, allowing the fermenters to achieve a steady state within the initial 8 days. The sampling period went from day 9 to 12. We evaluated the fermentation parameters of pH, volatile fatty acids, dry matter disappearance, digestibility, gas production, and methane synthesis.Zeolite did not affect methane production; it only numerically reduced methane, with methane concentration (g/L) decreasing from 32.33 (CON) to 28.99 and daily methane production (mg/g) from 18.84 to 16.82. Digestibility results showed no significant difference in dry matter (DM) or organic matter (OM) digestibility, and starch and neutral detergent fiber (NDF) digestibility remained unaffected. Crude protein (CP) digestibility was significantly reduced (P = 0.022) with ZC inclusion. No significant effect was found in VFA proportions or total production. Ammonia concentrations tended to increase (P = 0.080), and microbial nitrogen production tended to decrease (P = 0.056). Zeolite’s high affinity for NH4+ may have directed ammonium away from microbial protein synthesis, and the high NH4+ might explain the lack of reduction in methane production in zeolite diets
Effects of Leptin Signaling on Xenopus Tail Angiogenesis in Development and Regeneration
Get PDFThe role of leptin as a pro-angiogenic factor is well-established in multiple contexts in adult mammals, including wound healing. Additionally, leptin can act both at the local level of a wound and as a neuroendocrine signal. Prior research from our lab has shown that leptin increases the rate of regeneration in larval Xenopus laevis. However, it is currently unknown whether leptin acts as a pro-angiogenic factor during development and regeneration, and how leptin regulates regeneration via its different signaling pathways. This study demonstrates that leptin’s angiogenic role is conserved in peripheral tissues of developing and regenerating X. laevis, and explores multiple mechanisms through which leptin may promote regeneration and angiogenesis via local and neuroendocrine signaling pathways. In the first chapter, leptin is shown to promote peripheral angiogenesis during development. Leptin protein is expressed in and binds its receptor in developing blood vessels. Increased leptin signaling during the embryo-larval transition increases tail fin vessel length and development, while decreased leptin levels have an opposing effect and cause disorganization of tail vasculature and decrease of angiogenic gene expression. The second chapter demonstrates that leptin also promotes vascular regeneration. Local leptin expression increases regeneration rate and growth of regeneratingblood vessels, while decreasing leptin reduces these. Gene expression analysis suggests that leptin regulates regeneration by promoting cell survival, inflammation, cell motility, ECM remodeling, metabolism, and cell proliferation. Leptin administration increases expression of genes that promote angiogenesis at the tissue and blood vessel level. Finally, the third chapter describes differences in vascular regeneration and gene expression due to local and systemic (i.e., neuroendocrine) leptin administration. Both physiological measurements and gene expression data show that both types of leptin administration accelerate tail regeneration and rate of angiogenesis into the blastema, but increased local leptin signaling caused an earlier increase in regeneration rate and vascular regeneration, while increased neuroendocrine leptin signaling via IP leptin injection had a later effect. Additionally, the function of gene clusters were similar between local and neuroendocrine leptin signaling, but the timing of this expression was affected by application method. This dissertation shows that leptin’s pro-angiogenic actions are conserved in models of development and regeneration, and that it modulates gene expression in multiple pro-regenerative pathways and affects the timing of responses based on application method. These results may have applications in the fields of wound treatment and regenerative medicine, identifying leptin as a potential therapeutic for angiogenic wound healing and a master regulator of development, wound healing, and regeneration
MANA MOʻOLELO THE POWER OF STORY AND ITS CREATION
Get PDFThis dissertation is built upon ancestral ʻike (knowledge), Kanaka Maoli scholars such as Mary Kawena Pukui and Dr. Manu Meyer (Pukui, 1983; Pukui et al., 1979a, 1979b; Meyer, 1998, 2001), and KanakaCrit (Reyes, 2017). Research on representation in young adult literature is added to amplify what Kānaka Maoli always understood, “Words can heal; words can destroy” (Pukui, 1983). Through this foundation and a Kanaka Maoli methodology called Māʻawe Pono (Kahakalau, 2019) the writing of a modern moʻolelo (story) for young adults was conducted. The moʻolelo and the cultural process of creation was analyzed using ʻōlelo noʻeau (proverbs) and the research by Brandy McDougall (2016) whose work offers insight on what is a moʻolelo. Findings offer a window into how a modern Kanaka Maoli can do culturally relevant work, where the research process doesn’t strip away their identity but builds on it leading to a decolonizing experience
EARLY WILDFIRE DETECTION VIA UAVS MANAGING DATA SCARCITY AND RESOURCE CONSTRAINTS
Get PDFIn recent times, climate change is an increasingly important environmental issue. One of the side effects of rising temperature is a significant increase in wildfires. Theresulting wildfires cause a significant amount of damage ranging from direct property damage to decreases in air quality. Due to the unexpected nature of the event, wildfiremanagement is a difficult task. The best method for managing wildfires is through early wildfire detection. Traditional methods for early wildfire detection rely on eitherfire watchtowers or satellites. However, these methods suffer in either the temporal or spatial resolution.Recently, the use of Unmanned Aerial Vehicles (UAVs) combined with on-board deep learning has received a surge of interest for early wildfire detection. Compared to traditional methods, UAVs have improved spatial and temporal resolution while being cheaper to implement. However, current research into the solution still suffers in several key aspects. The deep learning can be categorized into training, where a model is trained to detect wildfires, and inference, where the model is deployed on-board. For the training process, data collection and generalization of models have been identified as key issues. To address the issue of data collection, the use of One Class Classification is proposed. Various One Class Models were evaluated on the FLAME dataset and achieved an accuracy of 92%. For inference, the on-board nature of the application makes resource constraints a major design problem. In particular, energy usage, memory usage and inference speed are important metrics to consider. To improve the energy usage and inference speed of an on-board model, a method using sequential pixel variation is proposed. The method was evaluated on the FLAME dataset and achieved significant reductions in energy usage and inference speed while maintaining similar accuracy
OPTIMIZATION OF N-TYPE CDTE AND CD-SE-TE FOR NEXT-GENERATION SOLAR CELLS
Get PDFThis dissertation establishes a comprehensive framework for advancing cadmium telluride (CdTe) photovoltaics through a systematic exploration of n-type absorber design, defect control, surface passivation, thin-film epitaxy, and device simulation. While conventional p-type CdTe has reached a practical efficiency plateau, this work demonstrates that iodine-doped n-type CdTe (CdTe:I) and indium-doped Cd-Se-Te (CST:In) can overcome intrinsic limitations. Using thermoelectric effect spectroscopy (TEES), Hall-effect, photoluminescence (PL), and density functional theory (DFT), the fundamental defect levels of CdTe:I were identified, revealing shallow donors (ITe ~0.05 eV) and compensating acceptors (VCd, ITe–VCd) that govern Fermi-level pinning. Post-growth Cd annealing was shown to suppress compensation centers, activate iodine donors, and reduce resistivity, enabling high-quality n-type material.At the interface level, Br passivation most effectively reduced surface recombination, while CdCl₂ improved bulk lifetime, highlighting their distinct chemical roles. Thin films fabricated by close-spaced sublimation epitaxy (CSSE) confirmed epitaxial growth of CdTe:In, CdTe:I, and CST:In, and preliminary homojunction devices established the feasibility of all-CdTe n-type structures. Further, SCAPS-1D simulations predicted efficiencies >25% for optimized MXene/p-CdTe:As/n-CdTe:I/In designs, while prototype devices revealed performance losses from mobility degradation and interface recombination. Extending to ternary alloys, CST:In crystals exhibited nearly 100% dopant activation, electron concentration up to 9.5×10¹⁸ cm⁻³, mobility >600 cm²/V·s, and lifetimes approaching the radiative limit—demonstrating a new high-performance absorber platform without intensive post-growth treatment.
Overall, this dissertation pioneers the transition from p-type to n-type CdTe absorber technologies, integrates bulk single-crystal and thin-film approaches, and provides both mechanistic insights and design principles for next-generation CdTe-based solar cells
STRUCTURAL PERFORMANCE OF SIMPLY SUPPORTED CROSS-LAMINATED TIMBER DIAPHRAGMS
Get PDFCross-laminated timber (CLT) has emerged as a prominent mass timber product, enabling the construction of mid- and high-rise buildings with improved sustainability and efficiency. A critical component in such structures is the floor diaphragm, which is responsible for transferring lateral loads induced by wind and earthquakes to the vertical resisting elements and for defining the overall lateral behavior of the system. This response is strongly governed by its in-plane stiffness. Despite its importance, the behavior of CLT diaphragms is not yet fully understood, and current design provisions remain limited.This dissertation investigates the in-plane performance of simply supported CLT diaphragms through comprehensive numerical analyses. The research focuses on the influence of key parameters, including aspect ratio, splice connection stiffness, and the presence of boundary elements such as chords and collectors. Each parameter was examined individually through parametric studies and subsequently in combination to capture interaction effects. The numerical models were calibrated against available experimental data and implemented in a systematic framework to evaluate both stress distributions within the diaphragm and global displacement patterns.Results demonstrate that the stiffness and overall behavior of simply supported CLT diaphragms depend strongly on the stiffness of splice connections, their orientation relative to the applied lateral load, as well as both global and local aspect ratios. Furthermore, boundary elements such as chords and collectors were shown to effectively limit relative displacements between panels and, in some cases, alter the distribution of stresses within the diaphragm. This redistribution may lead to stress concentrations that are not typically considered in current design approaches. The analyses also revealed that as diaphragm stiffness increases, the structural response shifts from an I-beam type mechanism, with stress concentrations at the diaphragm ends, toward a deep-beam behavior where stresses are distributed more uniformly across the diaphragm.The analyses further demonstrated that relying on traditional simplified analytical approaches, such as those commonly applied to timber light-frame systems, does not always provide a realistic prediction of CLT diaphragm behavior. These methods do not account for the fact that, depending on connection stiffness, the diaphragm may act as a single unit or as a series of beams. Moreover, they neglect potential stress concentrations around splice regions, which can be critical for accurate design assessment.The primary contribution of this research is the development of design-oriented recommendations for simply supported CLT diaphragms, addressing gaps in current standards. These recommendations are framed within a performance-based design philosophy rather than a purely strength-based approach, explicitly accounting for factors such as splice stiffness, panel orientation relative to the applied load, global and local aspect ratios, and the presence of boundary elements. This work enhances the understanding of CLT diaphragm behavior and supports the advancement of performance-based timber engineering
Human-Centered Automation in Software Engineering
Get PDFThe rapid expansion in scale and complexity of software systems over the past two decades has necessitated the development of advanced software engineering (SE) technologies aimed at enhancing developer productivity and reducing maintenance costs. A central challenge in this evolution is ensuring that these automated systems align with human factors—developer perceptions, needs, and workflows—to be effective and widely adopted. Misalignment can lead to tools that, despite technical sophistication, fail to deliver practical benefits to their intended users.This work focuses on bridging the gap between automated SE technologies and human factors, emphasizing the importance of a human-centric approach in the development and evaluation of these tools. We begin by examining how reliance on automatic evaluation metrics as proxies for human assessment can create a disconnect between SE technologies and developers. Specifically, we analyze two contexts: (a) code summarization, where automatic metrics are pivotal benchmarks yet may not reflect developer preferences, and (b) code readability assessment, where metrics intended for direct use by developers may not align with their perceptions of readability. Our findings reveal that commonly used automatic metrics often fail to accurately represent human assessments, underscoring the need for more human-aligned evaluation methods.Next, we explore how integrating insights from human factors research, such as program comprehension studies, can enhance the effectiveness of SE technologies. First, we propose human-centered interventions in automated testing: an underutilized domains with a wealth of literature that has not seen large scale adoption. By incorporating developer feedback and cognitive principles, the intervention is designed to align more closely with real-world workflows and have been preferred by practitioners over traditional methods, demonstrating the value of a user-guided development approach. Second, we extend our human-centric focus to applications of Large Language Models (LLMs) in software development tasks, which are increasingly being used across a diverse set of subdomains in SE.This work underscores the necessity of aligning automated SE technologies with human factors at every stage—from evaluation to system design. By placing human comprehension at the forefront of computation, this work contributes to the development of more effective, adoptable, and trustworthy tools in software engineering, ultimately bridging the divide between automated systems and their human users
YODA: WEARABLE HUMAN ACTIVITY DETECTION
Get PDFThe problem of Human Activity Detection (HAD) is to detect every occurrence of a target activity within a sensor time series. Unlike Human Activity Recognition (HAR) algorithms, which map a segment of a time series onto an activity label, HAD algorithms provide labels of activity occurrences together with the start and end time of each occurrence. HAR models have demonstrated high performance in controlled laboratory datasets, and their success highlights the importance of leveraging pretrained recognition models to enhance HAD systems.In this research, we introduce YODA (You Only Detect Activities), a unified framework that formulates HAD as a temporal object detection problem inspired by computer vision. We present the problem of activity detection and explain how the YODA algorithm processes sensor time series to detect target activities. We evaluate YODA on existing public datasets. Additionally, we further evaluate performance for a new dataset that we collect for individuals labeling activities in the wild.
To improve performance, we further consider improvements to YODA that build upon the strengths of existing HAR backbones. Experimental evaluations show that YODA effectively detects locomotion and static activities, while challenges remain for brief transitional actions. By integrating pretrained HAR knowledge into detection tasks, this work bridges the gap between recognition and detection. YODA offers a novel human activity detection method for real-world human activity analysis using wearable sensors
Biosolids as a Soil Amendment: Implications for Soil Health, Carbon Sequestration, and Sustainable Crop Productivity
Get PDFBiosolids are recycled wastewater products (sewage sludge) that have been treated to meet EPA standards for agricultural land application. They are primarily applied as alternatives to nitrogen (N) fertilizers but also contain organic matter (OM; 60-80% dry weight), carbon (C; 35-40%), and other plant essential micro- and macro-nutrients. Biosolids provide an avenue for waste recycling that returns nutrients and organic materials from food back into agricultural systems, thus, closing nutrient cycles, and potentially offsetting energy used in the production of synthetic N fertilizers. Biosolids also have the potential as OM and C sources to improve soil health, store additional soil C, and help farmers integrate additional soil health practices into rotation.The research presented in this dissertation was conducted in semi-arid dryland (rainfed) agricultural systems, which face specific challenges relating to moisture limitation and erosion. Thus, many of the research questions and objectives of this work are focused on how biosolids might positively impact soil health and C accrual within these systems.
Chapter two of this dissertation investigated the effects of biosolids on critical soil health functions for semi-arid systems in two long-term biosolids trials, one located in a Washington (WA) state grain-fallow system which compared three biosolids rates to a synthetic N fertilizer and unfertilized control, and the other in a Colorado (CO) system, comparing biosolids applied based on soil N availability to synthetic N fertilizer in two cropping rotations (wheat-corn-fallow and wheat-fallow). Application rate and timing impacted the effects of biosolids at the two sites. In WA increasing biosolids applications increased soil C and N pools, N-acetyl β-glucosaminidase, phosphomonoesterase enzyme activity, available water holding capacity, and decreased bulk density. At both sites, biosolids increased microbial biomass as measured by phospholipid fatty acids and Mehlich-3 extractable soil P. Treatment effects with biosolids at the CO site were within the wheat-corn-fallow rotation, which received the highest biosolids application of the two rotations. This study shows that biosolids can improve key soil health functions of OM and nutrient cycling, and water storage in semi-arid dryland cropping systems.
Chapter three of this dissertation investigated how long-term (28 year) applications of biosolids influenced soil organic C (OC) accumulation and stability by measuring soil OC stocks and fractions (water extractable OC, free particulate OC, occluded particulate OC, and mineral associated OC). The trial was located on a long-term WA biosolids trial, with three biosolids application rates (4.9, 7.0 and 10.0 Mg ha-1), applied every four years, compared to synthetic fertilizer and an unfertilized control. C stocks were measured on soil cores collected as deep as possible (typically ~60 cm), and cut in 15 cm segments (0-15, 15-30, 30-45, 45-60 cm). This study found that in semi-arid dryland systems with low clay content, the 7.0 and 10.0 Mg ha-1 rates of biosolids increase soil OC stocks and free and occluded particulate OC but not water extractable or mineral associated OC, indicating that the more stable fraction of C may not change with additional C inputs under current site conditions. However, this study does show that additional C can be stored in this system in particulate OC fractions with biosolids compared to synthetic N fertilizers.
Chapter four of this dissertation investigated the short-term (two applications, three years apart) impacts of biosolids applications and cover crop grazing on soil health, soil moisture, and yield, along with the influence of biosolids on grazable biomass in semi-arid systems. The experiment was a split plot design with cropping system as the main plot (cover cropping + grazing or chemical fallow) and fertility as the sub plot (biosolids, synthetic fertilizer, or an unfertilized control). This study found that short-term biosolids applications are effective at increasing grazable biomass and nutrient availability in semi-arid dryland systems and highlights the need for continued applications at agronomic rates for soil health benefits observed in long-term trials. Biosolids were also found to have the potential to reduce soil moisture due to increased biomass growth in biosolids amended plots but cover cropping vs fallowing did not seem to effect soil moisture availability