48440 research outputs found
Sort by
Ultrasound based Hand Motion and Force Tracking powered by Deep Learning for Human-Machine Interfacing
No full textWith the advancements in artificial intelligence and machine learning, there is a growing interest in human-machine interfacing systems. This is because now it’s possible to process large amounts of data to extract meaningful information. Biosignal based human-machine interfacing aims to utilize signals obtained from the human body to control digital, physical and augmented/virtual reality systems. Forearm ultrasound data is particularly attractive in this domain since it provides a visualization of the forearm musculature responsible for how the hands interact with the world. This dissertation focuses on utilizing machine learning to use the data provided by ultrasound to make meaningful estimations of hand movements. Specifically, the research focuses on estimating hand gestures, finger angles and finger forces. A major focus has been given to estimating multiple quantities in a single end-to-end pipelines for real-time human-machine interfacing. Towards that, simultaneously estimating finger angles and hand gestures has been explored. This kind of simultaneous estimation also finds relevance for robotic manipulation, towards which, simultaneously estimating manipulation skills and forces applied by hands on objects has been studied. Additionally, efforts have been made towards improving the design of the wearable ultrasound device, and improving the performance towards estimating these quantities. By advancing the capabilities of forearm ultrasound based hand movement and force estimation, this research is a step in the direction of utilizing a widespread non-ionizing medical imaging modality towards effective human-machine interfacing towards a future where seamless control of robots and digital and virtual interfaces becomes a necessity
Accelerating digital engineering and digital transformation within the U.S. defense industry using strategic adoption influencers
No full textThe U.S. Department of Defense has emphasized the critical need for digital engineering (DE) and digital transformation (DT) to maintain technological superiority in an increasingly complex global landscape. Despite strategic initiatives and mandates such as DoD Instruction 5000.97, adoption across the defense industry remains fragmented and inconsistent. This dissertation investigates the factors influencing the adoption of digital engineering and digital transformation technologies and innovations within the U.S. defense industry. Motivated by the observed gap between Department of Defense strategic intent and actual implementation, this research examines how "Strategic Adoption Influencers" (SAI) drive progress towards digital transformation. Through a comprehensive literature review and analysis of adoption theories and models, the study identifies key SAIs impacting the attainment of DE/DT strategic goals and objectives. Building upon this foundation, the study employs a survey-based methodology and deductive approach to analyze the perceived impact of various SAIs within the defense industry. This research offers practical guidance for United States (U.S.) defense organizations seeking to accelerate DE/DT adoption by enabling them to leverage key drivers and address specific challenges. The insights gained contribute to a deeper understanding of the complex dynamics influencing adoption of digital innovations leading to DE and DT. Furthermore, the findings offer valuable lessons for other sectors grappling with similar digital transformations
Desktop CNC Milling Machine
No full textThis project explores the development of an open-source desktop Computer Numerical Control (CNC) milling machine, including the integration of an improved control architecture and sensor suite, with the intention being to address the lack of accessible desktop-machining options for students and hobbyists. Industrial CNC machines cost over 2,800, provides users with an intuitive graphical interface for controlling the machine and the additional functionality; the design also gives users the ability to manufacture the machine with minimal outsourcing. The modular base-model supports a coolant system, automatic tool changer, 4th-axis, and mechanical bed leveling. User friendly and open-source software were used and customized to expand the capabilities of the previous three axis system. The machine additionally included many electrical safety systems for regular use in classroom settings. Furthermore, various sensors were utilized during the development, producing data that validated the effectiveness of all its functionalities. The machine produced repeatability within 0.005”, acceleration within ±0.03 g, safe motor temperatures. The final result was an affordable CNC mill that allows users to become familiarized with the basics of machining in a safe environment
Preserving a Dying Artform: On-line Cataloguing Tool for Cibachrome Artworks
No full textCibachrome, a photographic printing process using silver bleach dyes, began production in the 1950s and was discontinued in 2012, making it hard to locate and identify. The goal of our project was to assist the Cibachrome Association in developing an online collective catalogue prototype for sharing Cibachrome. Through case studies and interviews with artists, collectors, conservators, professors, and restorers, we found that inaccessibility of information, lack of community, issues with inaccuracies in online information, and lack of interest in current platforms keeps the Cibachrome community scattered. After developing the prototype, our team recommended implementing visual tools, increasing engagement to aid connectivity, ensuring accurate and standardized information, and site guides
Digitizing Plan for the Fusilier Museum London
No full textThe Fusilier Museum London plans to release digital reproductions of their World War II collection in time for the 80th anniversary of the 1945 Allied Victory in Europe. Our team was tasked with creating a digitization plan that the museum will use to facilitate their digitization efforts. We identified best practices for digitization and scoped the 16,726 pages of the WWII collection. We surveyed 50 visitors and tracked visitor attendance, creating a dashboard to visualize trends in the data that could aid the planning of museum events. Through a careful evaluation of digitization companies, we found that Max Communications Ltd. would best fulfill the needs of the museum, largely due to their affordable pricing and prior experience with regimental museums
Mathematical Finance Project 2024/25
No full textEvery investor desires a cost-efficient portfolio which achieves a specific value distribution with a minimum initial investment. Understanding cost-efficiency in discrete incomplete markets is critical for financial modeling. Dybvig proved that in discrete complete markets, the lowest cost method to realize an investment distribution follows reverse state price density (price per probability) order. While foundational results have been established for complete market systems, those in which all risks can be hedged, Bernard and Sturm extended the cost-efficient payoff calculation to incomplete markets with a 3-state model using a risk neutral measure per probability in place of a state price density. The risk neutral measure per probability is not unique. As a result, the highest minimum of the cost-efficient payment found from a family of risk neutral measures is the worst-case scenario. Bernard and Sturm find that the convexified maximin, maximin, and convexified minimax of the cost-efficient payment are all equal for the 3-state model while the minimax is different. This study generalizes the Bernard and Sturm 3-state model to n-states, deriving the convexified maximin, maximin, convexified minimax, and minimax of the cost-efficient payment for the n-state system. Simulated annealing solves optimizations. Comparing with earlier 3-state results verifies the code correctness for convexified maximin, maximin, convexified minimax, and minimax. The code correctness is further confirmed by comparison with an explicit 4-state solution which is independent of the n-state generalization. For both 3-state and 4-state cases, the convexified maximin, maximin, and convexified minimax give the same result. The equality of the convexified maximin, maximin, and convexified minimax of the cost-efficient payment for both 3 and 4-state cases demonstrate that the notion of cost-efficiency from complete markets remains well-defined. This contribution paves the way for future research into optimal portfolio selection for n-state systems in incomplete markets
PRIMO: Mobile 3D Printer
No full textThe 4 billion market on its own. Current commercial solutions have limited build heights and massive logistic requirements for larger machines. This project created a prototype mobile robot capable of 3D-printing cement while driving atop previously printed layers to build structures without a limit on height. The robot’s small form factor enables it to drive atop previous layers and reduces logistics requirements. A fleet of these robots could be deployed to a remote location without the infrastructure that a larger machine would require. In space applications, sending many of these robots is cheaper and more redundant than a large machine currently on the commercial market. The scope for this MQP was limited to printing straight walls with continuous printing, even during material refilling
Comparing Oxygen and Radical-Initiated Hydrothermal Liquefaction for Polystyrene Recycling
No full textPlastic pollution is an environmental crisis exacerbated by inadequate recycling methods, particularly for polystyrene. This MQP explores hydrothermal liquefaction with hydrogen peroxide, a chemical recycling method, as an alternative for traditional recycling methods. In order to understand the role hydrogen peroxide plays in the depolymerization of polystyrene, traditional hydrothermal liquefaction, radical-initiated hydrothermal liquefaction (RI-HTL) with hydrogen peroxide, and oxygenated hydrothermal liquefaction (O2-HTL) are compared. At fixed conditions, RI-HTL provides the highest yields of the three at roughly 88%, followed by O2-HTL at roughly 20%, and finally traditional HTL at roughly 8%. GC-MS and FTIR reveal that O2-HTL produces a high number of oxygenated compounds, followed by RI-HTL, and then traditional HTL. These results indicate that hydrogen peroxide plays both a radical initiating and oxidative role in the depolymerization of polystyrene. RI-HTL shows promise in providing a less severe way to chemically recycle polystyrene into high value products
Development of a Papillary Muscle Adjustor for Atrioventricular Valve Simulators
No full textHeart valve regurgitation, particularly involving the mitral valve, is a common cardiovascular condition caused in part by papillary muscle displacement. In vitro heart simulators provide a safe, controlled environment for studying such conditions and testing potential medical interventions. However, accurate modeling of regurgitation requires precise and adjustable control of papillary muscle positioning. This project aimed to develop a papillary muscle adjustor capable of displacing muscle tissue in three degrees of freedom—apical/basal, anterior/posterior, and lateral/septal—within a left heart simulator. Through iterative design, CAD modeling, and 3D printing, a functional device was produced that demonstrated controlled displacement and operational durability in water-based heart simulations. The adjustor was tested in a porcine mitral valve model and successfully integrated into an in vitro heart system, replicating both healthy valve function and valve stenosis. Although full regurgitation testing was not completed due to equipment limitations, the device lays a strong foundation for future studies and offers an accessible, reproducible tool for cardiovascular research
Structural Studies of EGFR/Kek1 binding: The Role of Ligand
No full textThe Epidermal Growth Factor Receptor (EGFR) family are receptor tyrosine kinases whose dysregulation is implicated in numerous cancers due to their role in cell proliferation, survival, and differentiation. Among them, ErbB2 remains in a constitutively active conformation and lacks a known ligand, making it a frequent driver of oncogenesis. While targeted therapies like Herceptin inhibit ErbB2 dimerization, therapeutic resistance has prompted the search for novel strategies. In Drosophila, dEGFR adopts an extended configuration similar to ErbB2 and can be inhibited by a transmembrane protein called Kekkon1 (Kek1). This could involve interfering with receptor dimerization or ligand binding, but the precise mechanism remains unknown. Notably, Kek1 does not bind the human EGFR family, suggesting structural barriers to cross-species therapeutic use. To explore Kek1’s mechanism and therapeutic potential, I sought to investigate how the presence of the activating ligand, Spitz, alters dEGFR–Kek1 interactions. To do this, I expressed and quantified two molecular variants of Spitz, created and expressed two dEGFR variants, C-S and ∆C, in the Fc tag to prevent Fc tag–mediated dimerization, and demonstrated Spitz and dEGFR-C-S variants are functional in our in vitro binding assay. With these tools in hand, ongoing work aims to determine whether Kek1 competitively inhibits Spitz binding, if Kek1 can bind independently of ligand, and whether the timing of ligand addition influences receptor–inhibitor complex formation. These insights could clarify Kek1’s role as a structural inhibitor and inform future design of therapeutics targeting ligand-independent EGFR activation in cancer