136879 research outputs found
Sort by
Search for New Resonances Produced by Bottom Fermion Fusion in Dimuon Final State with Run 2 CMS Data
No full textA search is presented for a new neutrally-charged boson Z��� decaying into two muons associated with one or two jets and at least one b tag among them. The analysis is performed using data corresponding to an integrated luminosity of 137.58 f b^���1 collected in 2016-2018 with the CMS detector in proton-proton collisions at ��� s = 13 TeVand sets limits upon Z��� masses between 125 and 350 GeV. No significant deviation from the expected background is observed, setting 95% confidence level upper limits on the production cross section times branching ratio times acceptance (�� �� Br(Z ��� ��� ����)�� A) ranging between 3-10 f b. Acceptances for a variety of final states are also provided, allowing for the interpretation of these results to a variety of models
Characterization of a Hypersonic Turbulent Boundary Layer Using the VENOM Technique
No full textModeling fluctuating flow parameters of high speed laminar and turbulent boundary layers is essential for understanding the thermal loading experienced by vehicles re-entering the atmosphere. Hypersonic turbulent flows in particular are challenging to model due to the strong coupling of velocity and energy which drive heat flux. Experimental data of the relationship between these fluctuating quantities are limited due to the required simultaneous measurements of velocity, temperature, and density. Laser diagnostics have become key to measuring these fundamental parameters in unsteady hypersonic flows due to their non-intrusive nature and ability to provide fluctuating measurements, which can be used to assess current predictive turbulence models.
Mean and instantaneous velocity and temperature measurements using the ���Invisible Ink��� Vibrationally Excited Nitric Oxide Monitoring (VENOM) laser diagnostic are presented to characterize the hypersonic boundary layer above a 2.75 degree half-angle wedge test article. The measurements were performed in the Actively Controlled Expansion (ACE) blow-down wind tunnel under both laminar and turbulent conditions. A double dependent Gaussian fitting velocimetry algorithm was developed to account for laser reflections at the wall, and intensity fluctuations due to turbulent motion. Additionally, rotational temperature profiles were using a thermometry algorithm in conjunction with velocity information extracted from each image. The results demonstrated a uniform laminar flow across the flat plate which broke down to turbulence in response to inserted mechanical trips. Freestream fluctuations compare favorably to previous measurements. The laminar boundary layer fluctuations peak around 12% for the velocimetry results and 20% for the thermometry results. The turbulence fluctuations peaked around 18% and 24% for the velocimetry and thermometry results respectively. Using the laminar results as a baseline for the technique uncertainty, it is estimated the true turbulence fluctuations are on the order of 10���15%. The mean and instantaneous measurements qualitatively agree with previous separate velocity and temperature measurements and simulations. The generated database of instantaneous parameters can be used to directly determine the variable heat flux of the flow over the wedge surface using a constant specific heat model to assess current predictive turbulence models. The current measurements are analysis can also serve as a basis for future multi-component velocity and temperature boundary layer measurements, contributing to a comprehensive understanding of turbulence, a complex 3D phenomenon
Sequential Decision-Making Under Uncertainty in Multi-Robot Target Tracking Application
No full textA significant advance in robotics and automation is making robots handle uncertainty, a prerequisite to deploying robots outside the lab. To interact with a stochastic environment, the significant effort of evaluating possible conditions challenges the robotic system���s reaction and functionality. For example, how to meet the typical 1 Hz decision frequency with online planning to accommodate uncertainties in a multi-robot system, combined by robots maintaining multiple robotic layers in a single-board computer.
This thesis focuses on such trade-offs between computation and optimality in the world of uncertainty. We formalize planning tasks as Partial Observation Markov Decision Process (POMDP) as it models uncertainties in a horizon-based mathematical model.
The first topic is to plan under time-dependent stochastic constraints, specifically traffic signals. The vehicle���s pass at the intersection depends on the stochastic fixed-pattern traffic signal, introducing the time dimension to the planning space. We propose a spatial horizon-based search to avoid such a curse of additional dimensionality in the planning. Furthermore, the uncertainty of the future signal is overcome by the event of an initial signal change, after which the problem is no longer stochastic. With probabilistic modeling of traffic signal dynamics, we define such planning as an MDP and obtain the solution.
Uncertainty needs tackling in partially observable environments and multi-agent systems. Our second topic, multi-sensor active target tracking, involves interacting with objects of interest (OOI) in a multi-robot system. The objective is to deploy multiple robots with a limited field of view (FoV) to maximize the system-wise target tracking performance. We formulate the problem as a POMDP and apply the approximate dynamic programming (ADP) method to generate the receding horizon control policy for the POMDP. Planning horizon incorporates the estimation of OOIs to handle tricky target-tracking scenarios such as divergent OOI trajectories. On the other hand, extending the horizon increases resilience to the uncertainty of the world model, such as the unknown object of occlusion to the sensing.
The multi-agent aspect injects other robots��� actions, which become a consideration in the uncertainty of the ego agent. Agent-by-agent planning reduces the complexity of planning, and we found the performance boundary of such sequential decision-making based on the locally Greedy algorithm. Since the original locally Greedy algorithm does not consider agents��� policies, which are not updated, adding the intention of those agents shows improvement based on our empirical studies. From implementing multi-agent planning based on receding horizon and objective optimization, we observe the emergent behaviors of cooperation. We show the consistency between planning algorithms��� optimality and the frequency of cooperative behavior in active target-tracking scenarios.
In summary, this dissertation mainly includes algorithmic research in robotic planning in an environment with various uncertainties: sensing, world model, and other agents��� decisions. It also contributes to the eco-driving and multi-robot sensing system application in the problem statement, algorithm design, behavior analysis, and simulation
Improving the Ability of Activity Recognition Systems to Detect Activities of Daily Living Performed In-the-Wild
No full textFailing to keep track of the performance of activities of daily living (ADLs) can lead to adverse health outcomes for people with health concerns. However, current recommended practices for keeping track are tedious and burdensome, making it easy for people to forget or stop managing their health. Using activity recognition systems to automatically detect and record ADL performance would address this issue, but most works in activity recognition focus on controlled or semi-naturalistic data in contrast to real world, in-the-wild data. As such, real world ADL recognition remains an open problem. Specifically, real world ADL recognition requires tackling several fundamental challenges for machine learning systems, and it is unclear if existing approaches would be robust to these challenges. We expect that semi-naturalistic data does not capture the diversity of all of the everyday activities such a system would encounter and that robust performance requires using in-the-wild data.
In this work, we focus on quantifying the challenges associated with in-the-wild settings and investigating the design of in-the-wild ADL recognition systems. To achieve these goals, we conduct a series of analyses and machine learning experiments on two ADL datasets, one semi-naturalistic and one in-the-wild. First, we measure the class imbalance, interpersonal variability, and pairwise class overlap to motivate the difficulty of recognizing in-the-wild data. Second, we demonstrate the importance of training on negative samples, showing that training on NULL data results in more robust models than using unknown class rejection. Third, we investigate the design of in-the-wild ADL recognition systems, exploring both classical and deep learning methods as well as models with varying levels of context of the user���s hands. In doing so, we develop a recognition system that can recognize several ADLs with high event-based recall and precision with only the context of the dominant hand. These efforts represent a thorough investigation of a challenging open problem in human activity recognition. The results and insights serve as a meaningful step forward toward making robust in-the-wild ADL recognition a reality in order to make it easier for people to manage their health
Alternating Direction Method of Multipliers for Convex Optimization Problems with Tensorflow Using GPU Computing
No full textIn aerospace engineering, there are many high-dimensional optimization problems, including sensor optimization, control variable calculation, structure analysis, and more. Solving these high-dimension optimization problems demands substantial computational resources, leading to high computational costs. Mathematical optimization algorithms and computation methods have long been a focal point of study, aiming to reduce the complexity of high-dimensional problems and minimize computation costs. This research applies the optimization algorithm, alternating direction method of multipliers (ADMM), to multiple convex problems and provides an implementation of these algorithms in Python using TensorFlow for accelerated graphics processing unit (GPU) computing. While an existing public domain implementation of ADMM is available in MATLAB, this research strives to handle high-dimensional problems more effectively than this MATLAB implementation, which relies on central processing unit (CPU) computing. The TensorFlow implementation for GPU computing is then compared to the CPU computing implementation and the performance of the interior point method (IPM) when solving the same problem
How Are Standards Developed? An Analysis of the Science Standards Revision Process in Texas
No full textThis study aimed to determine how science education policy is developed and the expertise employed throughout the process. Texas was selected as a historical case due to the recent revision of the science standards, their large impact, and documentation availability. A historical perspective informed the research questions: 1) What are the key positions, underlying justifications, and factors that contributed to the development of science standards in the state of Texas? 2) What expertise do official and unofficial actors possess in the science standards revision process? 3) To what extent were Schwab���s commonplaces (subject matter, teachers, learners, and milieu) considered during the revision process for science standards? and 4) What is the State Board of Education review process followed when revising science standards? State Board of Education meetings were analyzed using naturalistic inquiry and within-case methods.
Findings indicate that several points of failure occurred throughout the revision process that may affect the viability and implementation of the resulting standards. These points of failure included flawed process design, transparency issues, and limited representation of expertise domains. Notably, the learner was not represented during the revision process. As a result, the standards risk being developmentally inappropriate for students; initial evidence suggests this is the case for several standards. To ensure standards are scientifically accurate, developmentally appropriate, can be implemented by teachers, and respectful of the environment in which it is enacted, expertise from the subject matter, the learner/learning, the teacher, and the milieu must be equally considered. Additionally, research, development, and diffusion (RD&D) should be utilized to ensure the standards and curriculum materials are appropriately tested before being required for teachers and students
An Epidemiologic Investigation on Arsenic Exposure and Its Association with Mental Health Outcomes
No full textMental health problems, characterized by a clinically important disturbance in cognitive, emotional, or behavioral aspects of an individual���s life, have garnered growing interest in the field of epidemiology due to their complex causal mechanisms, increasing prevalence, and substantial burden on modern populations. There are many causal factors suspected to impact mental health, mental illness, and emotional wellbeing including certain environmental toxins, like arsenic, although these associations are not well established. Arsenic has been shown to be a globally toxic substance that is associated with numerous health problems. Recent studies suggest that arsenic may have an impact on mental health due to its neurotoxicity.
The present research investigates the possible association between arsenic and mental health problems through a systematic review, geographically weighted regression on a county level for the 48 contiguous states with soil arsenic, and a geographically weighted regression on a county level for Texas with water arsenic exposure estimates. The systematic review demonstrated associations between Arsenic and multiple mental disorders as well as generally poor mental health. The first geographically weighted regression of the contiguous 48 states found that soil arsenic was a significant predictor of both county suicide rates and the years of potential life lost due to suicide rate. The second geographically weighted regression found that estimated water arsenic exposure was a significant predictor of both county suicide rates and number of days reported with poor mental health.
The present research suggests that arsenic may be a potential environmental risk factor for mental health outcomes. Further prospective studies with individual-level data are necessary to confirm causality and explore the interactions with other risk factors
On Fatalism: Revitalizing Durkheim's Theory in the Current Milieu
No full textEmile Durkheim���s fatalistic form of suicide serves as the sociological theory necessary for understanding contemporary social problems with the imbalance of fatalism in society. This has been disastrous throughout the twenty-first century. Specifically, the newer and more virtual the generation, the greater the impact of fatalism as the consequence of modernity from the twentieth century. Fatalism has been neglected or even omitted in both Durkheimian studies and social research on generations. Recontextualizing Durkheim���s theory in the current milieu serves to not only rediscover his works, but to apply them with relevance to the current social malaise that is afflicting society across demographic categories. The current social environment is excessively confining and enforces social behavior through undifferentiating bureaucracy. The increased dehumanization of the average person requires a comprehensive approach to social thought, exemplified by Durkheim and works that are connected to him. Specifically, this thesis draws on Durkheim���s opus, Suicide, and various other literature to study the current generations Millennial and Generation Z. Fatalism linkages impact: public health, suicide, autonomy, mental health, and other social problems. Productivity in sociological research relies on clear, substantive theoretical understanding and policy implications exist in addressing social problems by comprehending their commonalities and clarifying interdisciplinary perspectives
Generational Ethnography in Post-1974 Cyprus: Island Narratives and Shifting (Im)Mobility Through Nation-Making, Belonging, Identity, and Border Realities
No full textThe 1974 conflict and the troubles of the 1960s in Cyprus hold a unique position from an anthropological forced migration studies perspective. The Coup D���etat and the following Turkish intervention in 1974 was the tipping point of the inter-ethnic conflict in Cyprus, which led to thousands of Cypriots becoming refugees through the division of the island. Through extended interviews, this study looks at both Greek and Turkish Cypriot perceptions of the events before the 1974 conflict and the 1974 conflict, casting light on forced migration within the conflict. This study discusses how forced migration and ethnic cleansing divided the once-intermixed ethnic community. Additionally, this study looks at the transgenerational understanding of the ethnic, political, and geographical national divide of Cyprus through the dispersed communities. An understanding of the complex structure of the Cypriots from this study hopes to expedite the solutions to the issues of the current climate of the island
A Characterization of the Vertical Structure and Mixing of the Eastern Mediterranean Sea
No full textObservational data collected by THEMO (The Texas A&M - University of Haifa Eastern Mediterranean Observatory) is used to characterize the seasonal variability of the water-column to understand the vertical structure and vertical mixing of the Levantine Basin in the Eastern Mediterranean Sea. Previous studies have shown the marginal Eastern Mediterranean Sea is characterized by hypersaline waters, strong water-column stratification, and regular seasonal atmospheric patterns. THEMO is composed of two surface buoys located off the coast of Haifa, Israel: one shallow buoy (water depth of 125 m; 10 km from shore) in the coastal zone of the Levantine Basin, one deep buoy (water depth of 1430 m; 60 km from shore), and one StandAlone McLane Moored Profiler (MMP), which is deployed close to the deep buoy. These instruments and buoys provide in-situ near real-time subsurface physical oceanographic observations and atmospheric observations of the Eastern Mediterranean Sea. The objectives of this research are: 1) to characterize and identify the processes that force the seasonality of the surface mixed layer depth, 2) to quantify the mechanisms that drive the hydrographic variability of the upper (surface to 1400 m depth) water-column, and 3) to categorize the stability processes of the Eastern Mediterranean Sea upper water-column using Turner Angle (Tu). Results show that the mixed layer depth varies seasonality from 250 m in Winter to less than 85 m in non-Winter months and is correlated with relatively stronger winter wind speeds (average of 7.12 m/s (Winter) to 4 m/s (non-Winter)). The upper 100 m of the water-column also experiences warming of 2 degrees C during Winter due to enhanced downward mixing of warm surface water with cooler subsurface water. Statistical analysis using empirical orthogonal functional (EOF) analysis shows that outside Winter months, variance in the principal component time-series is most correlated with the geostrophic current direction and accounts for about 50% of the total variance of water-column temperature and salinity. Characterization of water-column stability shows stable waters (-45 degrees < Tu < 45 degrees), i.e., no overturning, are most likely to occur in the upper 200 m of the water-column, and double diffusive mixing through salt fingering occur in waters below the mixed layer depth. The depth of salt fingering and stability are similar to other areas of the deep Eastern Mediterranean Sea and mid-latitude marginal seas of the world ocean