Search

This thesis tackles the need for ultra-low power operation in power-constrained cryptographic computations. An example of such an application could be smartcards. One of the techniques which has proven to have the potential of rendering ultra-low power operation is ‘Adiabatic Logic Technique’. However, the adiabatic circuits has associated challenges due to high energy dissipation of the Power-Clock Generator (PCG) and complexity of the multi-phase power-clocking scheme. Energy efficiency of the adiabatic system is often degraded due to the high energy dissipation of the PCG. In this thesis, nstep charging strategy using tank capacitors is considered for the power-clock generation and several design rules and trade-offs between the circuit complexity and energy efficiency of the PCG using n-step charging circuits have been proposed. Since pipelining is inherent in adiabatic logic design, careful selection of architecture is essential, as otherwise overhead in terms of area and energy due to synchronization buffers is induced specifically, in the case of adiabatic designs using 4-phase power-clocking scheme. Several architectures for the Montgomery multiplier using adiabatic logic technique are implemented and compared. An architecture which constitutes an appropriate trade-off between energy efficiency and throughput is proposed along with its methodology. Also, a strategy to reduce the overhead due to synchronization buffers is proposed. A modification in the Montgomery multiplication algorithm is proposed. Furthermore, a problem due to the application of power-clock gating in cascade stages of adiabatic logic is identified. The problem degrades the energy savings that would otherwise be obtained by the application of power-clock gating. A solution to this problem is proposed. Cryptographic implementations also present an obvious target for Power Analysis Attacks (PAA). There are several existing secure adiabatic logic designs which are proposed as a countermeasure against PAA. Shortcomings of the existing logic designs are identified, and two novel secure adiabatic logic designs are proposed as the countermeasures against PAA and improvement over the existing logic designs.

Applied systems thinking has rapidly developed through successive waves of development, and the current reigning paradigm is the revisioned approach to critical systems thinking. This research scrutinizes systemic intervention. It employs the methods of second-order science to apply some of its principles reflexively back on to the domain to discover two gaps: one between the espoused aims of systemic intervention and the adequacy of its methods, the other about its dependence on dialogic rationality. It also delves into its philosophical underpinnings to trace the reason for this gap to the 'ghosts' of rationalism. This is because modern Western thinking equates consciousness with intentionality. I argue that there is another well-recognised mode of consciousness, that of non-intentionality. I name these two modes as the becoming-striving and the being-abiding orientations. To address the gap, firstly, a characterisation of the systemic ontology is attempted. Three basic features are identified: mindful interconnectedness, enactive cognition and teleonomy. I also describe plausible political, epistemic and pragmatic goals for systems thinking arising from this ontology. Four methods from adjacent disciplines are examined in detail to show that these address the systemic ontology in better fashion than existing systemic approaches. These mature global contemporary approaches access knowings corresponding to the being-abiding orientation, absent in systems thinking. A suitable ontoepistemology for systemic knowing must comprise of two ontologies and epistemologies corresponding to each of the two consciousness modes: four component elements. Suitable conceptual models from other disciplines serve the purpose of these four components. Thus, a model of immersive systemic knowing is assembled, which meets the requirements of a framework for systems thinking in terms of the goals posited. A key feature of this research is the espousal of experiential knowing: not in a phenomenological sense, but in terms of a radical empiricism. It argues for the value of practical knowings that go beyond rationalistic formulation, which are always held in the margins (in the language of boundaries). Systemists must actively seek such experiential knowing to enact truly creative improvement. The only answer to the problem of knowing the world better is to know the shadow aspects of the knowledge generating system. This requires truly radical methods and an extended epistemology, all shown to be available plentifully in other practices and cultures. Testimony is provided from two field projects that were a part of these inquiries, and from practitioner accounts.

Imaging sonar systems have become increasingly popular in numerous applications associated with underwater imaging. Though multibeam sonar systems have been used in a variety of applications, the cost of these systems limits their use. The reason for the high costs has been identified to the use of large number of hydrophone array elements and hence large number of associated analogue channels and analogue-to-digital converters (ADC) that are required in high resolution imaging. In this thesis, an imaging sonar system has been developed with as few as four array elements to minimise cost. The inter-element spacing between any two array elements was chosen to be much greater than half the wavelength. In order to avoid phase ambiguity associated with wide array element spacing, the time difference of arrival is determined. Hence, for this purpose a wideband chirp signal was used. The return signals were divided into range cells to determine the target range. The time difference of arrival was obtained by correlating the range cells. Using the time difference of arrival, the direction of arrival (DOA) angle was calculated. The image of the target being illuminated was formed using the calculated range and the DOA values. The image pixel intensity at any pixel position was determined from the correlation result between the range cells. A simulation model was built to test the theory developed. Simulations were performed for various inter-element spacing and for four different target profiles types. Two objective metrics (signal to noise (SNR) ratio and peak signal to noise (PSNR) ratio) and a subjective metric (Structural Similarity (SSIM) index) were used to determine the performance of the algorithm and image quality. Image formed from the simulations using two hydrophone elements showed the presence of artefacts in the form of correlation sidelobes. The SNR metric showed a low gain of -5dB on comparison against a test image. PSNR and SSIM ratio showed a constant image quality over all the array spacing. The number of array elements was increased and linear operation like averaging was applied. The results showed no improvement in the gain and image quality. ii To overcome the problem of correlation sidelobes, a non-linear combining process has been proposed. Using the non-linear combining process it was found that the SNR showed an average gain of 10 dB on simulated data over the images formed without it. The PSNR and SSIM also showed a small increase in the image quality. The computational complexity of the proposed non-linear combining process was calculated by determining the number of multiplications and additions. The time taken to perform these operations on a SHARC ADSP 21261 chip was calculated theoretically. The calculations showed the feasibility of using this algorithm on a digital signal processing (DSP) hardware. An experimental prototype was built and performance was tested during sea trials. The data obtained was processed using a computer. The experimental results verified that the processing algorithm was effective in a practical system.

The focus of this dissertation is to examine the strategic responses of Indian room air-conditioner (RAC) manufacturers to changes in the business environment, specifically those brought about by economic liberalisation, market reform, and new foreign competition post 1991. Strategic management literature is replete with appeals for examining developing economy industries in a state of transition, yet very little data exist pertaining to the strategic responses of firms as they adjust to increasing competitive pressures, both domestic and foreign, within the Indian market. Using an emerging market- strategic factor - corporate strategy conceptual model developed by the researcher from strategic management literature, this research investigates how Indian RAC firms have been responding to the changes brought about by liberalisation and reform. Information was gathered by means of semi-structured interviews with respondents at three different levels within the domestic room air-conditioner industry. These included industry experts, RAC manufacturers, and dealers of RACs. In addition, secondary data was also collected in order to examine the strategic behaviour of RAC manufacturers over the past fifteen years, thereby providing a contextual/historical background to the primary data. The strategic analysis undertaken of the Indian RAC industry and the capture of rich data, obtained from the in-depth interviews in particular, make a major contribution to strategic management knowledge - specifically within an emerging market context such as India. The research findings show that liberalisation and reforms have increased foreign and domestic competition within the Indian RAC industry and substantially raised consumer demand levels. The results demonstrate that the institutional environment, and the RAC industry structure and market settings have had a significant impact on domestic organisations' strategy. Firm-specific resources and network/relationship capabilities, however, have exerted a smaller influence since the onset of economic reforms. Yet, the importance of the resource and relational perspectives can beseen in Indian RAC firms' responses to the changing environment. The research noted cost reductions, an expansion in RAC product ranges and marketing initiatives, and strategic alliances with industry participants as typical responses.

Metal additive manufacturing (AM) is an emerging technology for producing parts with virtually no constraint on the geometry. AM builds a part by depositing materials in a layer-by-layer fashion. Despite the benefits in several critical applications, quality issues are one of the primary concerns for the widespread adoption of metal AM. Addressing these issues starts with a better understanding of the underlying physics and includes monitoring and controlling the process in a real-world manufacturing environment. Digital Twins (DTs) are virtual representations of physical systems that enable fast and accurate decision-making. DTs rely on Artificial Intelligence (AI) to process complex information from multiple sources in a manufacturing system at multiple levels. This information typically comes from partially known process physics, in-situ sensor data, and ex-situ quality measurements for a metal AM process. Most current AI models cannot handle ill-structured information from metal AM. Thus, this work proposes three novel machine-learning methods for improving the quality of metal AM processes. These methods enable DTs to control quality in several processes, including laser powder bed fusion (LPBF) and additive friction stir deposition (AFSD). The proposed three methods are as follows 1. Process improvement requires mapping the process parameters with ex-situ quality measurements. These mappings often tend to be non-stationary, with limited experimental data. This work utilizes a novel Deep Gaussian Process-based Bayesian optimization (DGP-SI-BO) method for sequential process design. DGP can model non-stationarity better than a traditional Gaussian Process (GP), but it is challenging for BO. The proposed DGP-SI-BO provides a bagging procedure for acquisition function with a DGP surrogate model inferred via Stochastic Imputation (SI). For a fixed time budget, the proposed method gives 10% better quality for the LPBF process than the widely used BO method while being three times faster than the state-of-the-art method. 2. For metal AM, the process physics information is usually in the form of Partial Differential Equations (PDEs). Though the PDEs, along with in-situ data, can be handled through Physics-informed Neural Networks (PINNs), the activation function in NNs is traditionally not designed to handle multi-scale PDEs. This work proposes a novel activation function Self-scalable tanh (Stan) function for PINNs. The proposed activation function modifies the traditional tanh function. Stan function is smooth, non-saturating, and has a trainable parameter. It can allow an easy flow of gradients and enable systematic scaling of the input-output mapping during training. Apart from solving the heat transfer equations for LPBF and AFSD, this work provides applications in areas including quantum physics and solid and fluid mechanics. Stan function also accelerates notoriously hard and ill-posed inverse discovery of process physics. 3. PDE-based simulations typically need to be much faster for in-situ process control. This work proposes to use a Fourier Neural Operator (FNO) for instantaneous predictions (1000 times speed up) of quality in metal AM. FNO is a data-driven method that maps the process parameters with a high dimensional quality tensor (like thermal distribution in LPBF). Training the FNO with simulated data from PINN ensures a quick response to alter the course of the manufacturing process. Once trained, a DT can readily deploy the model for real-time process monitoring. The proposed methods combine complex information to provide reliable machine-learning models and improve understanding of metal AM processes. Though these models can be independent, they complement each other to build DTs and achieve quality assurance in metal AM.

The rapid expansion of wireless devices and services is pushing the limits of the current RF spectrum, necessitating innovative approaches to spectrum utilization. Traditional methods, which involve dividing the spectrum into fixed segments, are becoming increasingly inefficient for next-generation wireless systems. This talk explores the concept of spectrum awareness in these networks, focusing on how it can transform spectrum utilization, regulation, and security.My research involves developing a comprehensive understanding of specific applications, breaking them down into requirements for sensing, signal processing, and system-level considerations. This approach has led to the creation of hardware systems and signal processing tools that are foundational for spectrum-aware next-generation networks. My PhD thesis work demonstrates the practical applications of these tools, showcasing their ability to improve spectrum usage, enhance regulatory frameworks, and strengthen wireless security in real-world scenarios.

Given the role of distributional semantics in child language acquisition, adult linguistic development, and the conceptualization of abstract entities, the present investigation seeks to explore if the variable frequency of linguistic utterances across clusters of users in a social network can be correlated with divergent interpretations of an ostensibly shared concept within a discursive community. Are differing rates of linguistic entrenchment within a community a marker of divergent conceptualization? To begin to address this question, this present pilot study details the socio-cognitive processes underlying entrenchment of language and concepts, and develops a method for studying divergent conceptualization in the online social media network Twitter.

Combustion at high pressures has the potential to increase the efficiency and usability of jet engines in the presence of high moisture. However, not a lot of research has been done in the domain of high pressure combustion owing to various challenges of performing experiments as well as simulations in these conditions. The studies that exist mostly explore the combustion of fuels such as methane and its hydrates. Also, which chemical reactions are the most dominating reactions in these cases hasn't been explored much. In this work, combustion of a proper jet engine fuel, JP-10 is analyzed under high pressures in the presence of dry air as well as moisture by plotting temperature and axial velocity profiles in a 1-dimensional cross-flow diffusion flame setting using Cantera's open source code. It is found that increasing pressure increases the extinction limit for the fixed amount of JP-10 being burnt in air. This finding proves that for given a high pressure device, the combustion of jet fuel can take place even if it contains higher amounts of moisture, hence, increasing its application range and usefulness, especially in extreme environmental situations. To determine the most dominant chemical reactions, sensitivity analyses are carried out for the various cases of pressure for both dry fuel as well as with moisture, near the extinction limits of the corresponding pressures. Hence, showing that an increase in pressure plays a significant role in altering the pathway that the combustion chemical reaction is following, proving that the role of the chemistry of the fuel cannot be ignored while developing high pressure combustion devices.

The rise of NGS, big data, and ‘-omics’ has ushered biology into a new age, with the power to fundamentally change how research is approached. Rather than using a singular hypothesis, we can now incorporate more data-driven methods that drive new biological insights, explain emergent biological phenomena, and/or derive novel functionality. This thesis highlights the changing role of computation to both learn more about biological systems as well as leveraging data-intensive computational techniques to create new proteins and enzymes. The ability for computational approaches to drive biological understanding is presented in three studies. First, the laboratory evolution of DNA polymerases, the workhorses of replication, towards novel functionality is explored. In the three polymerases created, modeling and large scale approaches are used to demonstrate the additional capability of each new enzyme. Next, two independent studies in the genomic adaptations needed for E. coli cells to adapt a 21st amino acid (selenocysteine and nitrotyrosine) are presented. Next generation sequencing is used to better understand the mechanisms of how cells accommodate the increased fitness burden placed by an orthogonal translation system. Lastly, community-wide changes in the oral microbiome are studied in the progression towards periodontitis, with implications towards potential therapeutic targets. The capstone of this thesis leverages big data techniques to engineer novel proteins, the chief functional units within cells. Protein structural data is implemented into a convolutional neural network to associate amino acids with neighboring chemical microenvironments at state-of-the-art accuracy. This algorithm enables identification of gain-of-function mutations, and subsequent experiments confirm substantive improvements in stability-associated phenotypes in vivo across three diverse proteins. This work is the first demonstration of using deep learning to empirically improve protein function and opens a new avenue for protein engineering.

The discovery in the late 1990s that as many as 70 million people in Bangladesh were exposed to naturally occurring arsenic through their drinking water sources sparked widespread efforts to mitigate the risks. The intensity of these efforts declined after the mid-2000s and the limited monitoring done since has raised concerns about their long-term sustainability. The most recent Multiple Indicator Cluster Survey conducted by the Government of Bangladesh and UNICEF indicated that in 2012-2013 over 40 million people still drank water that tested above the World Health Organization (WHO) guideline for arsenic in drinking water, showing that it remains a major public health concern. This dissertation aims to elucidate barriers to reducing arsenic exposure in Bangladesh, pilot strategies to address them, and generate specific recommendations for the broad range of water supply stakeholders. The research was performed by researchers at the University of Michigan and Asia Arsenic Network, a water focused NGO in Bangladesh. The researchers worked closely with a diverse range of stakeholders to define problems and design study approaches. Field data collection was done in Phulsara and Goga unions in the southwest of Bangladesh. We investigated the state of arsenic mitigation efforts in the two study unions through an assessment of existing water supply infrastructure and surveys of rural households to understand their perceptions and behaviors regarding drinking water. The results indicate a need for better water supply planning, with mitigation strategies currently in place failing due to crumbling community water infrastructure and low levels of adoption of safe water practices. The effect of a low-cost (< USD 9/household) informational intervention on reducing arsenic exposure in arsenic affected rural households was evaluated through a randomized control trial study. The intervention consisted of sharing arsenic awareness messaging, an individual household water quality test result, and specific recommendations for alternate sources with improved water quality. The results show that the intervention led to a significant number of households changing water sources, thereby lowering arsenic exposure. This work highlights the benefit of continued well xv testing and educational programming in Bangladesh, efforts that have declined sharply since 2006. While the majority of community operated water systems fail within three years of installation, our work with the user communities of two such systems, has identified initiatives that can improve their sustainability. We documented these cases in a short video, which is intended to inform and motivate other rural communities to better manage their water infrastructure. The research further included a critical assessment of eight commercially available arsenic field test kits. While arsenic test kits are widely used in Bangladesh, an up-to-date assessment of their accuracy was lacking in the scientific literature. The results of this study show that several test kits, including the one currently most commonly used in Bangladesh, can provide variable results and often significantly underestimate arsenic levels. Finally, specific recommendations based on the work performed in this dissertation are provided for a range of water supply stakeholders in Bangladesh including government agencies, donor organizations, non-governmental organizations, and field kit manufacturers. The findings from this work are intended to draw attention to the continuing need for additional arsenic mitigation efforts and help the wide range of water supply stakeholders in Bangladesh make more informed decisions in their work.

Aging is a natural part of an organism’s life cycle. It is, however, associated with increased risks for many chronic diseases including obesity and sarcopenia. Aging is also associated with dampened functions of the circadian clock, an internal timing system that orchestrates physiological function and behavior and displays recurrent daily 24h rhythms. The circadian clock has an intricate relationship with metabolic regulators and plays a key role in the daily partitioning of energy producing/consuming processes for the efficient functioning of metabolism. Time-restricted feeding (TRF) – a feeding regime that restricts caloric intake to an 8-9 hour feeding window in the active phase of an organism – is conceptualized as a method of synchronizing feeding-fasting cycles with endogenous circadian clock. TRF is a particularly innovative dieting strategy because it does not require alterations in caloric quantity/quality, thus rendering TRF an easy to adopt and yet highly efficacious intervention. The ability of TRF to holistically improve numerous health parameters in young mice makes it an attractive candidate for the prevention (or delay) of sarcopenia and age-related metabolic decline. Here we showed that TRF prevents body weight gain, reduces fat mass, preserves lean mass, bolsters glucose regulation, improves endurance and strength, and enhances muscle clock gene expression, all without caloric restriction and irrespective of age in male C57BL/6J mice fed a western diet. Thus, TRF can delay age-related physiological decline in middle-aged mice and our results set stage for further exploration of the benefits of TRF across lifespan.

To feed a growing world population with limited amount of arable land, we must develop new meth- ods of sustainable farming that maintain or increase yield while minimizing chemical inputs such as fertilizers, herbicides, and pesticides. Precision agriculture techniques seek to address this chal- lenge by monitoring key indicators of crop health and targeting timely treatment only to plants or infested areas that need it. Such monitoring is still, often, a time consuming and expensive activity and thus not performed as standard practice. Developing automated methods for such monitoring using unmanned aerial and ground vehicles can thus provide a major impetus for the adoption of precision agriculture practices at scale. This thesis deals with improving the perception capabilities of autonomous systems that can be deployed for environmental monitoring, especially in agricul- tural contexts. Deploying autonomous systems on agricultural fields is a challenging task, since the perception system must deal with an unstructured, dynamic environment under widely varying illumination and weather conditions. This thesis makes contributions towards three parts of the perception pipeline required for most autonomous systems - calibration, mapping and inference. In the first part we look at radiometrically calibrating vision sensors, i.e. cameras in a field context. In contrast to existing, laboratory based approaches requiring specialized and expensive equipment, we develop a practical and modular method to radiometrically calibrate monochrome, colour and hyperspectral cameras with data that can be collected on the field. Our data driven, parameter free, maximum likelihood estimation based approach allows robust estimation of the sensor response, lens vignetting and global illuminant with minimal prior knowledge about the camera and lens setup in use. In the second part, leveraging developments in 3D photogrammetry, we propose methods to metrically map and extract plant trait indicators from crop fields using a low cost Unmanned Aerial Vehicle (UAV) carrying a camera. We show that crop parameters extracted using these 3D maps compare reasonably well with measurements taken on the ground and hence can be used to estimate the crop status, thus providing a practical and effective means of monitoring fields at high spatial and temporal resolution with minimal user intervention. Additionally, we propose a method by which a Unmanned Ground Vehicle (UGV) can localize itself in such a 3D map enabling higher resolution map updates and automated targeted interventions, such as delivering fertilizer or herbicide. We show that our collaborative mapping approach, which combines the 3D geometry of the crop field with a weakly semantic feature - a vegetation index, performs well on a wide variety of different crop fields and outperforms several state of the art map registration techniques for such scenarios. In the third part of this thesis, we present a generic, machine learning based framework for determining the types and severities of different stress factors affecting regions of a crop field using only remotely sensed data. Once trained, the classification models use radiometrically calibrated hyperspectral imagery, along with plant trait indicators extracted from a 3D point cloud of part of a crop field for quickly and systematically inferring the state of the crops with respect to the presence and severity of commonly occurring stress factors, such as nutrient deficiency as well as weed pressure. We show the effectiveness of the trained models by testing their predictions at different crop growth stages through a comprehensive greenhouse experiment following the growth of sugarbeet plants exposed to a variety of combinations of these stress factors. In summary, this thesis focuses on enabling autonomous systems to create and maintain quantita- tive spatio-temporal spectral representations of the environment and using these representations for pertinent inference tasks within the context of precision agriculture. Most of the contributions of this thesis are accompanied by open source software and datasets to enable the community to benefit from and expand on our work.

Entity normalization is an essential but challenging task for knowledge base construction by text mining the scientific literature. Related to entity linking and word sense disambiguation, models for entity normalization usually depend either on the surface text phrases of the entities or their coherence in the context. In this paper, we show that NormCo, a deep neural network normalization model, can switch between phrase and coherence models. Specifically, we tested this model on the tasks of normalizing bacteria and disease entities extracted from the scientific literature. These two entity types are important to construct a knowledge base of associations between diseases and human microbiome, an emerging development in biotechnology. We show that NormCo switched to either phrase or coherence model to accomplish the best performance for different entity types. We revised NormCo with a dynamic document-level switch and tested it with novel embedding techniques and obtained encouraging results. We organized and consolidated available lexical resources and annotated corpora for bacteria entity tagging and normalization, revealing a high level of discrepancy among these resources. Our results with these resources suggest that the skewed distribution of biomedical entity mentions may require different normalization approaches for highly mentioned entities from long-tail ones.

Online communities are online interaction spaces for people that break the barriers of time, space, and scale and provide opportunities for companionship and social support, information exchange, retail, and entertainment. Among them are online peer production communities that have a fantastic business model where volunteers come together to produce content and drive traffic to these sites. Although as a class these communities are successful, the success of individual communities greatly varies. To become and remain successful, these communities must meet a number of challenges related to starting communities, retention of members, encouraging commitment, and contribution from their members, regulating the behavior of members and so on. This dissertation focuses on the specific challenge of newcomer retention and productivity in the context of online peer-production communities. Exploring three different communities with entirely different structures and compositions – MovieLens, GitHub, and Wikipedia and building upon prior work in this space, this dissertation offers a number of important predictors of retention and productivity of newcomers. First, this dissertation explores the value of early activity diversity in the presence of the amount of early activity as a predictor of newcomer retention. Second, this dissertation digs into more fundamental psychological traits of newcomers such as personality and presents findings on relationships between personality and newcomer retention, preferences, and productivity. Third, this dissertation explores and presents results on the relationship between community interactions (apart from norms, policies and rigid structures) and newcomer retention. Fourth, this dissertation studies and presents the effects of various kinds of prior experience of newcomers on retention and productivity in a new group they join. This dissertation concludes by offering a number of directions for future research.

Cardiovascular diseases have been the biggest killer in the United States for decades. Despite tremendous research and new innovations, these trends do not seem to be slowing down, with more than 1 in every 3 deaths related to cardiovascular disease in 2016. In the U.S. cardiovascular diseases affect close to 92.1 million people, causing 807,775 deaths each year. According to the American Heart Association statistics report update of 2016, 17.3 million deaths worldwide were attributed to cardiovascular diseases. Moreover, this number is estimated to rise to more than 23.6 million by the year 2030. In addition, U.S. healthcare cost associated with cardiovascular disease has been rising with a massive $316 billion of associated direct and indirect cost, which is estimated to rise to $918 billion by the year 2030. Moreover, as the average age of the U.S. population increases this trend will most likely stay the same, if not increase. Etiologically, lack of oxygen and blood to the heart tissue is one the major cause of this ischemic injury and scar formation. This happens through the buildup of plaque in the arteries which reduces the blood flow and when left untreated can shut down the blood supply to downstream tissue. Our current medical approaches to deal with ischemic injuries include blood thinners, stents, beta-blockers, and surgical cleaning of the plaque. Although, heart transplants are the most optimal treatment, unavailability of donor hearts still remains a major concern when dealing with this problem. While microRNAs were first discovered in early 1990s by Ambros in C. elegans it was not until mid-2000s that first cardiac microRNA was discovered and studied. Binding to the 3’UTR (untranslated region) and inhibiting expression of the gene is the canonical mode of action for microRNAs, however there have been reports of 5’UTR binding and other non-canonical modes of action. Recent work by Matkovich et al. 2013 even determined miRNAs regulating other cardiac miRNAs. In addition to their discovery, microRNAs were shown to be differentially regulated in diseased and non-diseased cases. In addition, work by Eulalio et al. 2012 was one of the first to show miRNAs inducing proliferation in cardiomyocytes and inducing regeneration of lost myocardium, post injury. In this work, we have furthered the understanding of miRNAs and cardiovascular disease (specifically myocardial infarction) by showing a significant proliferation in adult cardiomyocytes both ex-vivo and in-vivo, which reduces infarction size and improves heart function as early as one month post treatment. We also confirmed the effect across different species (mice, rats, and human) and showed that miR1825 can induce significant proliferation in all three species. Lastly, we project a very high translational potential for this work due to the nature of miRNAs as a therapeutic: readily available, longer shelf-life, and no prerequisite for an autologous donor.

The Java Modeling Language (JML) is a behavioral interface specific language designed to specify Java modules (which are Java classes and interfaces). Specifications are used to describe the intended functionality without considering the way it is implemented. In JML, if a user wants to write specifications for a Java file, he or she must undertake several steps. To help automate the process of creating annotations for method specifications, a tool Jmlspec was created. Jmlspec generated a file that refines the source file and has empty placeholders in which one can write specifications. Although Jmlspec worked with older versions of Java, it does not work with the current version of Java (Java 8). This thesis describes the implementation of a new version of the Jmlspec tool that is compatible with newest versions of Java. This tool will be more maintainable than the older version of Jmlspec and easier to extend.

Group psychotherapy has become a preferred modality of direct clinical psychological services offered by many University Counseling Centers (UCCs), primarily because of the effectiveness of group psychotherapy in addressing the unique developmental issues that college students experience. However, there is a need for empirically supported data to identify barriers to group psychotherapy among college students, particularly related to multicultural identity variables. The present study examined barriers that college students face when considering joining a psychotherapy group. Participants completed a demographic questionnaire and either the original or the modified version of the Barriers Scale, (Harris, 2012). Participants were divided into three groups, Black/African-American, White/Caucasian or Other Minorities (O.M.). Results showed that group psychotherapy was found to be one of the least desirable methods for distress management among all three groups. However, White college students showed greater willingness to enroll in group psychotherapy when more information was provided about the types of groups. Furthermore, O.M. and White college students endorsed a fear of being judged because of their experiences with people of other race/ethnicities. Results indicated that the fear of being judged negatively impacted White college students’ willingness to participate in group psychotherapy, but had no impact for O.M. students.

The railway is an important critical national infrastructure network in many western countries and by its nature is exposed to diverse threats ranging from extreme weather and climate change risks, man-made and malevolent activities. The complex nature of the interdependencies across the respective systems that make up a typical rail network can result in the propagation of network failures with serious consequences for the movement of trains, passengers, goods and services. Recent disruption events have shown how rapid and widescale the impact of failures can be on complex interdependent rail networks, and with the projection of increased risk from extreme weather and climate change events, there is a need for robust methodologies that can address this growing complexity with applicability to real-world networks. The aim of this thesis is therefore to develop models and methodologies for analysing the risk and resilience of large scale interdependent rail infrastructure networks. It introduces multi-scale, system-of-systems based methodologies and applied analysis that provides important new insights into interdependent infrastructure network risk and resilience at national scale using real-world case studies. The thesis comprises three parts. First, a theoretical network-based model of the multi-track layout is built to replicate the actual operational functionality of the rail network which is tested using a real-world national timetable dataset for Great Britain. This model is calibrated using an empirical dataset of journey disruptions to arrive at a robust journey disruption network model. Next, by adopting a complex network-based modelling approach, real-world asset data is integrated from the respective systems of the national rail network in Great Britain to represent the physical interconnectivity that exists within and between its interdependent infrastructure systems. The coupling of both models give rise to an integrated infrastructure network model which is used to simulate scenarios of risk to the traction power network and resilience modelling. Lastly, the impact of extreme weather risks is studied with the development of a flood risk methodology which is tested by carrying out a comprehensive risk assessment of the national rail network in Great Britain. The criticality of individual assets is calculated by summing the direct impacts, with the indirect impacts measured in terms of train and passenger disruptions. The development of robust methodologies and research models that represent the actual physical, geographic and functional interdependencies across rail networks using real-world asset datasets allows for the testing of multiple scenarios of risk and resilience which can be used to support robust decision making in complex networks. This research therefore presents new theoretical insights and practical methodologies with direct applicability to academia, infrastructure managers, policy and decision makers across Great Britain and beyond.

A real time operating system (RTOS) is an important component of many embedded applications which are becoming more common in our day-to-day lives. These systems can be divided into soft or hard real time systems based on whether the timing parameters are flexible or not. Scheduling algorithms play a major role in contributing to the working of an RTOS, they are used to schedule the various tasks present in the system. It is therefore crucial for embedded system engineers to understand the working of various scheduling algorithms and the internal design structure of an RTOS.In this thesis, we design a task schedule generator using the C language. This schedule generator provides schedulability analysis for periodic, independent, preemptive tasks using Rate Monotonic scheduling, schedulability analysis and a schedule for periodic, independent, non-preemptive tasks using Earliest Deadline First scheduling, and a schedule for independent tasks using Priority scheduling. Further, we develop a variety of user functions that can help the user understand the internal working of our generator. This tool can be included in an RTOS that supports a variety of scheduling algorithms for a variety of functions. Further, the tool developed in this thesis can be used to study the internal workings of an RTOS schedule generator.

This dissertation examines the transient characteristics in next generation power electronic circuits at both the device-level and the systems-level. At the device-level, the effect of the parasitic capacitances on the switching performance of emerging wide bandgap semiconductors (WBG) is evaluated. Equivalent device models based on gallium nitride (GaN) and silicon carbide (SiC) are implemented in SaberRD and MATLAB, and transient switching characteristics are analyzed in great detail. The effects of the parasitic capacitances on detrimental circuit behavior such as “overshoot,” “ringing,” and “false turn-on” are investigated. The modeled results are supplemented and validated with experimental characterization of the devices in various power conversion circuits. The models can be used to aid in the design of next generation WBG devices so that the undesirable transient effects displayed by contemporary versions of these devices can be mitigated. At the systems-level, the transient overshoot demonstrated by conventional maximum power point tracking algorithms for photovoltaic power conversion systems is investigated. An adaptive controller is implemented so that the operating point can converge to the optimal power point rapidly with minimal overshoot. This new controller overcomes the parasitic components inherent to the power converter which limit its ability to deliver maximum power rapidly. It will be shown that with the new controller, the maximum power point is attainable in 4 milliseconds. The work accomplished in this dissertation lays a foundation for power electronic engineers to integrate semiconductor device theory with control theory to optimize the performance of next generation power conversion systems.

The generation of induced pluripotent stem cells (iPSCs) from somatic cells has opened new doors for regenerative medicine by overcoming the ethical concerns surrounding embryonic stem (ES) cell research. However, iPSC technology presents several safety concerns, such as the potential risk of tumor formation, that have caused apprehension for use of iPSCs in humans. One such approach is "direct reprogramming" which can bypass the iPSC or pluripotent stage and obtain tissue-specific cell types from somatic cells. In this study, we examined whether an important transcription factor involved in myogenesis (Pax3) or neurogenesis (NeuroD1) alone can directly reprogram the mouse embryonic fibroblasts (MEFs) into myogenic or neurogenic lineages, respectively. In addition, we created fusion transcription factors (Pax3 or NeuroD1) with the potent MyoD transactivation domain (MDA) that could facilitate radical acceleration of reprogramming into the desired cell type through chromatin modification compared to wild-type factors. Here, we showed that Pax3 can reprogram MEFs towards a myogenic lineage and that MDA-Pax3 further enhances this myogenic reprogramming event. In addition, ectopic expression of NeuroD1 and MDA-NeuroD1 is able to induce neurogenic genes in MEFs, suggesting the partial neurogenic conversion of MEFs. Furthermore, we also showed that the ectopic expression of NeuroD1 but not MDA-NeuroD1 in myoblasts could suppress myogenic differentiation. These data suggest that single gene transduction such as Pax3 or NeuroD1 will become a feasible therapeutic approach for neuro- and muscle degenerative diseases, respectively.

The coalburg seam coal is an example of difficult to float bituminous coal. Laboratory tests conducted on coalburg flotation feed sample revealed recovery values around 28% with 15% product ash when using fuel oil as collector under natural pH conditions. A detailed study showed that increasing pH from natural value of 5.6 to 7.5 provided a significant improvement in recovery of approximately 32 absolute percentage points. The improvement is believed to be result of the release of humic acids from the surface and the dispersion of clay particles thereby leaving a more hydrophobic surface. Based on the tests conducted with various commercially available collectors, oleic acid was selected as a model collector for oxidized coals. Conventional flotation tests found an increase in combustible recovery of 10 absolute percentage points above the pH improvement using 4:1 blend of fuel oil and oleic acid. The problem of higher ash in conventional cell product due to entrainment was minimized by the use of wash water in a flotation column. A flotation concentrate containing less than 7.5% ash was produced while recovering around 75% of the combustible material. Further testing using fatty acids-fuel oil blend also showed evidence of a near 200% increase in flotation rate.

When systems and subsystems are put under external shocks and duress, they suffer physical and economic collapse. The ability of the system components to recover and operate at new stable production levels characterizes resilience. This research addresses the problem of estimating, quantifying and planning for resilience in interdependent systems, where interconnectedness adds to problem complexity. Interdependence drives the behavior of sectors before and after disruptions. Among other approaches this study concentrates on economic interdependence because it provides insights into other levels of interdependence. For sectors the normalized losses in economic outputs and demands are suitable metrics for measuring interdependent risk. As such the inoperability input-output model enterprise is employed and expanded in this study to provide a useful tool for measuring the cascading effects of disruptions across large-scale interdependent infrastructure systems. This research defines economic resilience for interdependent infrastructures as an "ability exhibited by such systems that allows them to recover productivity after a disruptive event in a desired time and/or with an acceptable cost". Through the dynamic interdependent risk model resilience for a disrupted infrastructure is quantified in terms of its average system functionality, maximum loss in functionality and the time to recovery, which make up a resilience estimation decision-space. Estimating such a decision-space through the dynamic model depends upon the estimation of the rate parameter in the model. This research proposes a new approach, based on dynamic data assimilation methods, for estimating the rate parameter and strengthening post-disaster resilience of economic systems. The solution to the data assimilation problem generates estimates for the rate of resilient recovery that reflects planning considerations interpreted as commodity substitutions, inventory management and incorporating redundancies. The research also presents a robust optimization based risk management approach for strengthening interdependent static resilience estimation. There is a paucity of research dealing with quantification and assessment of uncertainties in interdependency models. The focus here is more on the extreme bounds of event and data uncertainties. The deterministic optimization becomes a robust optimization problem when extremes of uncertainties are considered. Computationally tractable robust counterparts to nominal problems are presented here. Also presented in this research is a discrete event simulation based queuing model for studying multi-modal transportation systems with particular focus on inland waterway ports. Such models are used for impact analysis studies of inland port disruptions. They can be integrated with the resilience planning methodologies to develop a framework for large-scale interdependent risk and recovery analysis.

Recently, a novel methodology termed Loop-mediated isothermal amplification (LAMP) was reported as the preferred technique for rapid diagnostic testing of pathogens. The advantages of LAMP over older techniques are its economic viability, rapidity and its obviation of complex instruments. Reported here is an interdisciplinary approach between medical practices and engineering to implement an affordable diagnostic device which employs LAMP for detection of pathogens. LAMP involves the optical excitation and detection of a pathogen sample mixed with a fluorescent dye as it is heated and amplified over an hour. The device reported here consists of readily available components which heat, optically excite and detect a LAMP sample. Finally, the integration will graphically display the amplification of the LAMP sample as a function of time on a palm top computer by exploiting the ubiquitous 802.11 wireless standard. The diagnostic box implemented here and its supporting components accurately discriminated between positive (infected) and negative (not infected) LAMP samples of various pathogens in approximately one hour. These results were verified using the standard method of pathogen diagnosis termed electrophoresis. Due to the low cost and portability of the device reported here, it poses as a potential solution to the need for quality point of care diagnostic tools in developing countries.

Around 1.7 million people in United States are living with limb loss and annually 185,000 more undergo amputation. 70% to 80% of the limb loss patients experience chronic post-amputation pain for which present methods of treatment are ineffective in providing pain relief. Evidence has been gathering that electric stimulation of peripheral nerves can effectively treat neuropathic pain conditions like occipital neuralgia, postherpetic neuralgia, CRPS etc. A peripheral nerve stimulation device to treat post-amputation pain is being studied by NDI Medical, LLC and its commercial feasibility was evaluated via market research that involved assessing present treatment methods, estimating market potential and physician interviews etc. In spite of a big market potential ($15B), a need for an effective therapy and a demand from physicians, certain factors like inapplicability in diabetic patients, etc. work against this device’s commercial feasibility.

The thesis is broadly divided into three major parts. In the first part, thermal imaging is used to experimentally measure temperature and a numerical model is developed to predict the thermal heating of a sample exposed to 3rd generation synchrotron X-ray beams. In this study we have experimentally measured heating by the beam using infrared imaging and used these measurements along with a theoretical numerical model to help understand the heating process. Specifically, the temperature rise of 1mm and 2mm glass spheres (sample surrogates) exposed to an intense X-ray beam and cooled in a uniform flow of nitrogen gas at two different flow rates is modeled, analyzed, and experimentally tested in an actual synchrotron beam. The heat transfer including external convection and internal heat conduction, was theoretically modeled using computational fluid dynamics (CFD) to predict the temperature variation in the interior and on the surface of the sphere when subjected to the X-ray beam. In the next part a heated disk cooled in a uniform stream of air is numerically investigated. Finite Volume Method (FMV) is used to study the effect of orientation of disk and imposed boundary conditions on the local and average heat transfer coefficients. Also a Nusselt number (Nu) correlation is developed, in terms of Reynolds (Re) and Prandtl number (Pr), for predicting forced convective heat transfer over isothermal or isoflux circular disk geometry at low Reynolds numbers (Re) in the range 10 to 150.Three different orientations of the disk are studied, where the disk orientation with respect to the flow is: a) parallel; b) inclined at 45; c) normal. Finally, in the last part the biocrystal along with the mother liquor is modeled as a short cylinder (i.e. disk). Both local “spot” heating and full heating of the crystal are considered. Two analytical models are developed that depicts the temperature variation within the crystal. A 1-D analytical solution is developed by treating the sample to be an annular “thin” fin; solutions presented revealing the radial variation in temperature distribution caused by small versus large beam size. Next a two-dimensional analytical solution is reported for full heating of biocrystal showing the depth wise spatial temperature variation in the disc. Results are reported considering more realistic sample sizes, X-Ray fluxes and flux densities and exposures.

A mixed technology FPGA (MT-FPGA) extends the positive aspects of conventional FPGAs into the non-conventional technologies like photonic, MEMS and RF systems. The results obtained from the first generation MT-FPGA architecture showed that by proper integration of the non-conventional technologies with the programmable FPGA architecture, the benefits associated with conventional FPGAs could be extended to the development of mixed-technology systems. The four principal components of the MT-FPGA are the multi-technology logic cluster (MTLC), the multi-technology block (MTB), the switch block and the I/O block. In this thesis, we are concerned the development of a second generation I/O module for the MT-FPGA. An attempt has been made to design the new I/O with increased flexibility when compared to the first generation I/O module. Also a number of novel features like dedicated clocks for signal sampling, unity gain analog buffers for processing analog signals and tri-state inverters at the I/O output have been incorporated in the second generation I/O. This I/O module has been incorporated in the design of the second generation MT-FPGA which has been fabricated using TSMC 0.25 ìm deep sub-micron process technology. The results obtained by testing the I/O module in coordination with the other MT-FPGA components have been summarized. Also, as part of this thesis, we have developed a signal flow indicator tool to predict the direction of signal flow within the second generation MT-FPGA. This tool also reports any programming errors present in a particular bit-stream configuration. The results obtained by running this tool against many benchmarks are also summarized.

For simulations spanning multiple length scales at continuum level, I present a parallel Generalized Interpolation Material Point (GIMP) method developed by Bardenhagen and Kober in the Structured Adaptive Mesh Refinement Application environment for 3D. SAMRAI is used here as a platform for domain decomposition, and for multi-level refinement in the areas of interest, to improve the computational accuracy and to reduce the overall computational time. A GIMP algorithm for the treatment of contact between a rigid conical indenter and a deformable work piece has also been developed for 3D. An efficient multi-level refinement and domain decomposition technique for parallel computations has been developed and validated using FEM. Expected speed-ups and memory savings has been achieved. The contact algorithm developed for 3D is validated using FEM. GIMP 3D code developed here, in general works fine for large deformations but FEM (ABAQUS) quits early.