137091 research outputs found
Sort by
Adaptive Backscatter Communication Approach for Next-Generation Biotelemetry Applications
No full textThis thesis presents the design and implementation of a Proof of Concept (POC) system that adaptively distributes logic gate resources of an Application Specific Integrated Circuit (ASIC) between RFID backscattering and sensor data processing. This work is an extension for the application of digital logic in RFID backscattering systems. The system utilizes the electronic switching capabilities of CMOS transistors within logic gates to create impedance variations between high and low states, effectively altering the Radar Cross Section (RCS) and thereby modulating the backscattered field.
The enhancement detailed in this thesis is the adaptive distribution of logic gate resources based on received signal power at the reader. A POC system is described in our work, where the reader classifies the received power as low or sufficient. If the received power is low, this information can be encoded into the carrier wave. This carrier wave is incident on the tag, where the tag performs two important tasks: 1) The control circuit decodes the control input from the carrier wave (This part is not in the scope of this Thesis.). 2) The FPGA receives this input and adaptively adjusts the amount of logic gate resources. In addition, the signal incident on the tag is modulated via backscattering from the FPGA’s chip. Hence, the FPGA’s chip is also responsible for the backscattering. The primary focus of this work is to enhance the received backscattered signal power by dynamically modifying the amount of logic gate resources on the FPGA based on a given control input. We detail how information received at the FPGA is utilized to make these adaptive adjustments. We also present the algorithm used to program the FPGA to enable this functionality. The practical implications of our work are demonstrated through successful measurements, which indicate that higher logic gate utilization correlates directly with increased received power. Our results show up to a 7dB improvement in received power when logic gate utilization of the FPGA is increased from 30% to 100%. We have also provided results showing that the system adaptively changes the logic gate utilization without the need for reprogramming the FPGA, using internal and external control logic to adjust the logic gate resources dynamically. In summary, this thesis aims to develop a fully self-sufficient RFID system that facilitates real-time adjustments to backscattering logic gate resources in the ASIC/FPGA based on the strength of the signal received at the reader. The findings from this study provide a solid foundation for the development of an adaptive RFID ASIC system for biotelemetry applicationsM.S.Electrical and Computer Engineerin
Soft Wearable Acoustical Sensing in Disease Autodiagnoses
No full textThis dissertation explores the foundational role of acoustic stethoscopes in modern medicine, acknowledging their crucial contribution to diagnostic processes by allowing physicians to listen to internal body sounds. Despite their effectiveness, these stethoscopes are limited by the subjectivity and expertise of the clinician, necessitating a profound understanding of intricate sound nuances. Recognizing these challenges, the research introduces a groundbreaking approach to acoustic monitoring through the development of flexible and adaptable soft wearable stethoscopes. Exploring into creation methods, materials, mechanical design, and electronic configuration, the dissertation emphasizes wireless soft wearables equipped with acoustical sensing capabilities. These innovations, applicable across a broad spectrum of diseases, facilitate continuous, wireless, real-time auscultation. An intriguing aspect lies in the exploration of biometric security, utilizing body sounds as a unique identifier. By combining flexibility, adaptability, and technological acumen, this research pioneers a transformative leap in stethoscope evolution, offering continuous monitoring and self-diagnosis potential. Ultimately, the dissertation signifies a paradigm shift in patient care and diagnostic accuracy through the integration of machine learning and soft wearable technology.Ph.D.Electrical and Computer Engineerin
Gromov Wasserstein Framework and Survey
No full textIn biological research, analyzing experimental data for cell type classification is a challenging and time-consuming task. Gromov-Wasserstein optimal transport methods offer promising results, but they face scalability issues with large datasets and variable features. To address this, we propose a Python-based framework facilitating the comparison of Gromov-Wasserstein implementations. We review literature, outline methodology, and assess scalability and accuracy metrics using simulated data. Results highlight Sliced GW's efficiency and Sampled Sliced GW's promise compared to other variations. Our work underscores the need for scalable and accurate computational methods in biological research, with implications for machine learning and bioinformatics.UndergraduateComputer Scienc
Enhancing mmWave Network Connectivity and Edge Intelligence through Reconfigurable Intelligent Surfaces for Communication and RF Processing
No full textThe next generation of wireless networks needs to support massive device deployments and ensure robust, widespread connectivity, particularly for edge devices in various Internet of Things (IoT) applications. To achieve this goal, two primary requirements should be addressed: (1) maintaining strong and reliable signal connectivity for a large number of devices, and (2) enabling real-time, on-device data processing for edge devices.
For the first requirement, millimeter wave (mmWave) technology has been proposed as a solution to achieve ultra-high network throughput. However, inherent characteristics of mmWave signals, including limited propagation range and high penetration loss, restrict the effective coverage of mmWave networks. In particular, non-line-of-sight (NLoS) issues arise when obstacles block line-of-sight (LoS) paths between transmitters and receivers, further limiting connectivity. The second requirement, on-device data processing for edge devices, is driven by the need for data privacy, real-time processing, and reduced dependence on internet connectivity. Yet, edge devices often face challenges due to limited computational resources, especially for local machine learning (ML) inference tasks. This constraint is increasingly significant with the growing demand for artificial intelligence (AI) applications in IoT scenarios. In this dissertation, we investigate the use of reconfigurable intelligent surfaces (RISs) to address the two challenges above.
First, we investigate the optimization of signal coverage in mmWave networks using RISs, since RISs can be leveraged to establish alternative links in the absence of LoS links. To this end, a coverage analysis framework is proposed based on stochastic geometry to evaluate the performance limits of RIS-assisted links, particularly in terms of signal sensitivity to obstacles. This framework enables evaluation of coverage performance for both single-RIS and multi-RIS links with different types of RISs, including reflective RISs and transmissive-reflective RISs. Through this comparative analysis, the proposed framework can provide insights to assist RIS deployment strategies. Then multi-RIS deployment strategies are proposed to enhance signal coverage in indoor environments. Two scenarios are investigated: (1) obstacles are randomly distributed in the room, and only stochastic information about their presence is available; and (2) obstacles are located at fixed positions, with full information about their locations and sizes provided. In the first scenario, analytical results for optimal RIS placements are derived to maximize connection probability when room dimensions and the number of available RISs are given. In the second scenario, a gradient-descent-based algorithm is developed to optimize the locations and orientations of RISs, aiming to minimize NLoS regions and thereby enhance coverage reliability.
To address the second challenge of on-device data processing, we explore the possibility of offloading computations from digital processors to the radio frequency (RF) domain by using RISs. To be specific, RISs serve as a medium for over-the-air (OTA) computation, where the goal is to manipulate RF signals that carry sensing or communication data directly in the RF domain, enabling signal propagation to emulate specific mathematical operations. RISs enable OTA computation by adjusting the electromagnetic properties of each RIS element to control incoming RF signals and produce the desired outgoing signals. In this way, RIS-based OTA computation offers the potential to reduce memory and computational demands on edge devices. In this dissertation, RIS-based OTA computation is explored for both non-ML and ML tasks. For non-ML cases, a direction-of-arrival (DoA) estimator comprised of one transmissive intelligent surface (TIS) and two receive antennas is proposed. The signal from the target passes through the RIS and is received by the antennas, with the DoA estimated from the power ratio between received signals at two antennas. This estimator offers advantages of low computational and hardware complexity, as it relies solely on received power, unlike classic methods that require phase information. In ML cases, RIS-based RF neural networks are explored, where sequentially placed TISs are used to mimic neural network layers for ML inference in the RF domain. Two types of RIS-based RF neural networks are investigated: (1) RF fully connected (FC) layers, and (2) RF two-dimensional (2D) convolutional layers. For FC layers, TISs with 1-bit and 2-bit phase configurations are used to reduce hardware complexity compared to continuous phase configurations, and a training scheme for quantized complex-valued RF neural networks is proposed. For 2D convolutional layers, an RF convolutional layer using three TISs is introduced to perform 2D convolution in the RF domain. Both types of RF neural networks are validated through simulations.Ph.D.Electrical and Computer Engineerin
Characterization of Conformational Dynamics of Archaeal Protein Tyrosine Phosphatases
No full textProtein tyrosine phosphatases have long been studied in bacteria and eukaryotes for therapeutic purposes. Key driver for the varying catalytic rates are conformational dynamics of catalytic loops that change with the loop sequence. While PTPs are thought to originate in archaea, only few archaeal PTPs have been studied. Despite the low available data, the proteins have displayed uncharacteristic loop dynamics and sequence motifs. To this date, no computational studies of archaeal PTPs have been performed. This thesis employs molecular dynamics simulations of the proteins TkPTP and SsoPTP as well as an ancestral archaeal PTP called N64 to characterize the loop dynamics of archaeal PTPs. Of key interest are first the findings that the P-loop undergoes correlated conformational changes and second that the changed D-loop sequence is an adaptation to thermophilic environments. These findings cumulate in the conclusion that the archaeal PTPs contain different conformational dynamics than conventional PTPs. With the new characteristics of archaeal PTP conformational dynamics, these insights help understand the development of an important family of signaling proteins.M.S.Chemistry and Biochemistr
A Parametric Design Tool to Support Customized Adaptations of Assistive Technologies for People with Hand Impairments
No full textThis thesis aims to explore the feasibility of a proof-of-concept parametric software that empowers occupational therapists (OTs) in fabricating customized adaptations of assistive technologies (ATs) to effectively address the distinctive requirements of individuals with hand impairments. The platform enables OTs to modify pre-designed models of utensil grips, serving as examples of ATs, based on individuals' hand assessments and anthropometric data. We conducted observations and interviews to gain insights into the AT adaptation processes by OTs for individuals with hand impairments. Subsequently, using an iterative design process, we created an interface prototype that allows for customized adaptations of ATs and invited an OT to participate in the preliminary testing of the prototype. Furthermore, a usability test was carried out with OTs to achieve a comprehensive appraisal of the parametric design tool's practical application and to obtain insights to the physical ATs for the hand that were generated. Lastly, we provide an outline of the design requirements for parametric modeling tools that support OTs and other clinicians to design customized ATs for individuals with limited hand dexterity.M.S.Industrial Desig
CuPBoP-AMD: Extending CUDA to AMD Platforms
No full textCuPBoP is an open-source CUDA translator that works to extend CUDA language support to a variety of GPU platforms by translating CUDA programs at LLVM/NVVM IR level. As a subsection of CuPBoP, CuPBoP-AMD is a CUDA translator that makes use of the founding ideas of CuPBoP to extend CUDA support to AMD devices. We propose that by translating CUDA programs at LLVM IR level, CuPBoP-AMD is able to translate a larger range of HIP programs without programmer intervention than existing translators, such as HIPIFY. Preliminary results show that programs translated using CuPBoP-AMD have approximately the same performance as programs translated by existing translators, while potentially improving the coverage of CUDA features supported by the translator.Undergraduat
Addressing Task Forgetting in Federated Learning for Anomaly Detection
No full textFederated learning is a new branch of research in the field of machine learning which leads to many new opportunities. Its objective is to bring together several entities to merge their model together in a so-called global model. This merged model respects data confidentiality and makes it possible to unite similar models developed on more varied data and in different contexts. Continuous learning which consists of retraining a model on new data is key for updating models. However, prior work shows that there is a risk of forgetting tasks learned from past data when employing continuous learning. This leads to a key question: is the global model in federated learning able to be designed to reduce tendency for task forgetting? The present study investigates federated learning models integrated with core memory backup to determine impact on task forgetting scenarios. Further, these federated models are compared with centralized learning models. To accomplish this, the study uses an Artificial Neural Network dedicated to the classification of machinery fault anomalies from bearing failure data. Multiple anomaly distributions encountered in production were evaluated including clustered, periodic, progressive and random types. The results demonstrate enhanced performance of federated learning with on average 20% to 50% better accuracy than centralized learning depending on the anomaly distribution in continuous learning. The federated learning model with memory also is shown capability to improve the stability of results.M.S.Mechanical Engineerin
Metabolomic analysis of predator urine for enhancing oyster shell strength and ecological implications
No full textThis thesis explores the role of metabolites in blue crab urine that enhance oyster shell strength, with downstream opportunity to address the ecological and economic challenges posed by declining oyster (Crassostrea virginica) populations in the Gulf of Mexico gulf oyster. Gulf oysters are essential to Gulf Coast ecosystems, providing water filtration, carbon sequestration, and coastal protection while supporting a significant industry. However, populations have declined by up to 85% due to overfishing, habitat loss, pollution, disease, predation, and climate change.
Using integrated spectroscopic analyses and database matching, along with oyster growth assays, the study identified a metabolite in blue crab urine that promotes oyster shell strengthening. The study focused on the metabolite most important for oyster shell strengthening based on prior partial least squares regression analysis of mass spectrometric metabolomic data. 3-(1-methyl-1H-imidazol-4-yl) propanoic acid was identified as the primary contributor to this effect, with structurally similar molecules, 3-(1-methyl-1H-imidazol-2-yl) propanoic acid and 3-(1-methyl-1H-imidazol-5-yl) propanoic acid observed in aquarium oyster growth assays to induce similar levels of shell strengthening. This generalized strengthening response reflects oysters’ adaptive strategy as r-selected species, which respond broadly to predator-associated cues.
The findings suggest that applying these chemical cues in oyster hatcheries could help produce a robust oyster spat more resistant to predation. Additionally, the study highlights the increasing power of metabolomic techniques to identify natural products in complex biological samples. Future research should employ and enhance these methods to enable the rapid identification of unknown compounds, which could be applied to conservation efforts to assist in the restoration of populations such as gulf oyster populations.M.S.Biolog
Process development and process analytical technology integration for cell therapy manufacturing
No full textBiomanufacturing of cell therapies involves highly complex and labor-intensive processes, where the process parameters and biological variabilities can significantly influence product quality, reproducibility, and therapeutic efficacy of the products. The complexity and largely manual unit operations contribute to product variability and high cost. To address these manufacturing challenges, we designed a digital-twin-enabled closed-loop cell manufacturing platform with automation and feedback controls. This platform integrates process analytical technologies (PAT) to enable deeper process understanding and provide real-time control of process variables.
Specifically, we designed bench-scale bioreactors with automated sampling, at-line and in-line monitoring, digital twin-enabled media nutrients estimation, and feedback-controlled feeding capabilities. Human umbilical cord tissue-derived MSCs (CT-MSCs) and T cells were used as the example cell therapy product. At-line glucose and lactate monitoring confirmed the accuracy of the digital twin estimations. Spent media samples and detailed functional characterizations of the MSCs and T cells end-products generated from the automation-controlled bioreactor demonstrated that high expansion and functions of the MSCs and T cells were maintained in these closed-loop bioreactors. Real-time imaging with quantitative oblique back illumination microscopy showed high-resolution images of cells in-process in a dynamic 3D environment. Overall, the digital twin-enabled bioreactor platform reduced costs, labor, time, and, more importantly, perturbations; and could improve yield while maintaining the phenotype and quality of cell therapy products. Our integrated automation system provides a blueprint for multiplexed PAT integration, process optimization, feedback-controlled intelligent automation to enable the discovery, monitoring, and control of critical quality attributes and critical process parameters for cell therapy manufacturing.Ph.D.Bioengineerin