7903 research outputs found
Sort by
Revit WalaSee: An Auto-Generated In-Wall Revit Model
No full textThe Revit WalaSee is a software-hardware integration tool designed to make building renovations smarter and more sustainable. In the construction industry, rework and demolition make up a large part of global construction waste and carbon emissions, largely due to the lack of internal wall documentation in older buildings. Our project addresses this issue by developing the pipeline between the Walabot wall scanner, a handheld device that detects hidden elements behind drywall, and Autodesk Revit, a widely used platform for Building Information Modeling (BIM).With the Revit WalaSee, someone without extensive technical training could scan any wall and produce accurate 3D modelling data.
By collecting and storing scan data from the Walabot, our system uses an algorithmic approach to recognize and label interior components, and converts them into an IFC file type (a 3D modeling file). In the 3D modeling file, each object is identified with properties such as material, size, and location. This IFC file can then be imported into any 3D modeling software that supports IFC files, though this project focuses on Autodesk Revit. This allows builders and designers to work with a digital representation of the wall’s interior, which improves planning and reduces waste. The goal is to make renovation projects more accurate, less wasteful, and promote sustainable renovation practices
Variable Natural Language Translator for Video Games
No full textIn the field of video games, developers work to immerse players in a variety of ways. In one genre of video games, 2D exploration, or Metroidvania as its colloquially known, developers try to immerse players by letting them explore large worlds with the hope to overcome isolation. Often these worlds are sparse or dead worlds where the player is alone, but by exploring more of the world familiarity with the space is achieved and this isolation can be overcome. We hope to innovate in this genre by utilizing a new way to simulate and overcome this isolation, through language learning. The hope is to create a world that seems alien at first, as the in-game NPCs (non-player character) speak a fictional language, but opens to the player as they learn this new language.
We achieved this innovation using parsing tools and natural language processing to create a translator that can translate English dialogue into our fictional language at different learning levels. This allows the player to interact with the NPCs and the game at varying language levels, where the game shows words the player has learned at that level in English, with the unlearnt words still displayed in the fictional language. This allows the player to interact with the game and simulate the feeling of learning a new language and culture to overcome the feeling of isolation through knowledge and understanding
EMG Vocal Translations
No full textIn this paper, we propose a novel augmentative and alternative communication (AAC) framework for silent speech. Many individuals with speech impairments are unable to vocalize effectively due to various conditions that affect the vocal cords. To engage in social activities, many rely on AAC devices that often lack flexibility and expressiveness. Users may still find self-expression and spontaneity difficult with such devices. This project presents a novel approach to developing a silent speech interface (SSI), providing a more adaptable and user-centered solution to give the vocally impaired a voice. Using surface electromyography (sEMG) alongside machine learning techniques, we aim to map neuromuscular signals produced by sub-vocalizations to audible phonemes, the smallest unit of speech. Sub-vocalizations are described as inner speech, the process of silently pronouncing words in one’s mind while reading. Unlike most alternative and augmentative speech devices, which are constrained to a set of predetermined and commonly used words, our approach would be able to provide unrestricted vocabulary through the compositional construction of phonemes. Our framework utilizes non-invasive surface electromyography sensors placed on speech articulator muscle groups, so our model can learn associations between subtle myoelectric patterns and the phonemes that are produced. We conducted a user study to collect synchronized EMG and audio data. With the data from these participants, we trained our model to classify phonetic symbols from windows of EMG data
Information Theoretical Analysis of Deep Neural Networks
No full textWe are at the dawn of an interesting era in the history of humankind where artificial intelligence (AI) is starting to play a significant role in our daily lives. Large Language Models (LLMs) like ChatGPT, Bard, Llama and others have shown a remarkable ability to converse with humans in many different languages across myriad topics. Machines which “learn” do so by creating a network of artificial neurons which loosely mimic the neurons in the human brain. Even though humans are now increasingly interacting with these machines, a clear understanding of how these machines exactly learn, remains elusive. In this study we look under the hood to explore how information flows through the layers of the machine’s artificial neural network, some of which are several levels deep. This dissertation looks at deep feed-forward neural networks, Convolutional Neural Networks (CNNs) and Generative AI models through the lens of Information Theory. It uses information theoretical concepts such as Rényi’s generalized entropy (of which Shannon’s entropy is a special case), mutual information, information channel capacity, various f-divergence measures and information geometry on Riemannian manifolds to understand the inner workings of these deep neural networks. Information Theoretic Learning (ITL) techniques based on information particles and information potential are used in this thesis to illustrate how information theory can be used to improve the learning process in neural networks. This thesis also combines principles from physics and statistical mechanics with information theory to describe how a generative AI model can learn the latent parameters of the input data to generate new samples
2025 AIAA Design/Build/Fly
No full textThis thesis presents the design, fabrication, and testing of Santa Clara University\u27s 2025 AIAA Design/Build/Fly (DBF) competition aircraft. The project comprises two integrated systems: the Mothership (MS), a 6-ft wingspan electric aircraft, and the Autonomous Parasite Glider (APG), designed to complete four missions: rapid assembly, airworthiness, efficient fuel transport, and autonomous glider deployment. MS employs a tapered wing with an MH 20 airfoil and a Scorpion SII-4025-520 kV motor, selected for aerodynamic efficiency and mission-specific payloads, including wingmounted steel-filled fuel tanks. The APG, constructed from carbonreinforced foam, is deployed via spring latch and guided by an Ardupilot GPS controller.
Development followed an iterative process encompassing conceptual, preliminary, and detailed design phases, supported by custom and industry-standard tools including MATLAB™, SOLIDWORKS™, XFLR5™, and a custom mission simulator (LapSim). Ground and flight testing validated design objectives, confirming rapid modular assembly, structural robustness, and autonomous navigation. Results emphasize the value of early integration, iterative prototyping, and simulation-driven design. Future improvements are recommended in autonomy, mission systems, and scalability
Trigger Buddy
No full textGun violence in the United States, particularly unauthorized firearm use by children and impulsive suicides, remains a devastating public health crisis. Despite widespread use of gun safes and locks, many remain ineffective or inconvenient, with most unauthorized users accessing firearms that were either unlocked or secured with easily bypassed mechanisms. Biometric locks present a promising alternative, but existing products often fall short in speed, reliability, and usability. To address this gap, we developed Trigger Buddy, a biometric handgun lock designed to balance accessibility, durability, and security. Trigger Buddy mounts directly to a handgun\u27s picatinny rail and uses a fingerprint sensor to rapidly unlock sliding steel doors covering the trigger well. Through iterative design and testing, we engineered a modular system with a steel-reinforced composite door, ergonomic housing, and a rack-and-pinion locking mechanism, all powered by a microcontroller and 9V battery. Our experiments demonstrated that Trigger Buddy meets critical performance metrics: unlocking in under 1 second, withstanding 100 lbf of prying force, and showing no statistically significant impact on shooting accuracy. The fingerprint scanner maintained a false rejection rate under 2% (excluding pinky fingers), and the total system weight remained close to the 1 lb target. Finite element analyses and physical four-point bending tests confirmed structural integrity under realistic loads. Trigger Buddy offers a secure, rapid-access solution for firearm owners, especially parents concerned with child safety. Its modular design and integration with standard firearm interfaces allow for ease of adoption. However, reliability under environmental stress (e.g., wet or dirty fingers) and long-term durability require continued refinement before commercial release. To ensure widespread utility and ethical impact, we recommend future iterations improve biometric robustness and reduce cost through manufacturing optimizations. Pursuing California DOJ certification is critical for market entry. Ultimately, Trigger Buddy demonstrates how innovative engineering can address real societal issues, bridging the gap between security and accessibility in firearm safety
‘Immigrant Invaders,’ ‘Arab Terrorists,’ and ‘Leftist Subversives’:Surveillance State from Palestine to Mexico
No full textImmigrants in the US exist under a very strict regime of surveillance. This regime is characterized by authentication systems, check-in points, registration, forms of mobility tracking, interoperable databases, and, at the militarized US-Mexico border, the scrutiny of facial-recognition technology, automatic watchtowers, and drones. However, this didn’t come about spontaneously. This paper attempts to outline the development of the immigration surveillance state as a generations-long bipartisan project. From the evolution of the military-industrial complex and technologies like the internet innovated to spy on national independence movements, to the proliferation of data collection made for global finance, the immigration surveillance state reveals a dense web of private beneficiaries who dictate its operation and the narrative surrounding it. Its long-term trajectory also illustrates another pattern, wherein the practices, technologies, and economic policies shaping surveillance at home are overwhelmingly developed in the context of imperialism and colonial domination. Few struggles illuminate this as transparently as that of the Palestinians, who are subjected to technologies and policies of colonization that are imported by American police and ICE. Indeed, there is a line that travels from repression and apartheid abroad, to the surveillance of immigrants, and eventually the surveillance and repression of political dissidents at home, as is now being witnessed with the case of Palestinian organizer Mahmoud Khalil. The incoming Trump Administration, and its connections to Elon Musk and the private tech sector, project vicious escalations not only to the surveillance of immigrants, but also to the repression of mass movements threatening the capitalist hegemony. However, this paper urges readers to look beyond individuals, and put on trial the entire system which brought these actors to power, which tends toward surveillance to preserve control over its subjects, and which time and again prioritizes profit over people’s needs
Curriculum Preferences and Engagement of Online Entrepreneurship Students: The Influence of Age and Gender
No full textIt is important for instructors and institutions to create learning experiences that are engaging, effective, and meaningful for students. To achieve these goals, instructors must understand the preferences and interests of their students, build engaging lessons based on those interests, and mitigate content that might make students feel excluded. In-person learning allows instructors to gather information about interests and engagement through direct interaction with students. Gathering information about student interests and engagement is more difficult for asynchronous, self-paced, online training programs. In this paper, we assess the interests, engagement, and disengagement of learners accessing online content focused on entrepreneurship. We focus on the influence of the demographic variables of age and gender and utilize data from Google Analytics to test hypotheses about the relationship between these variables and how the students interact with elements of the curriculum. STATA 18.0 was used for the statistical analysis. We find that while there are significant differences in the preferences for different elements of the curriculum based on gender and age, there is no evidence of a significant difference in curriculum engagement based on these demographic factors. These results support the conclusion that entrepreneurship students in this kind of learning environment select topics that are influenced by their gender and age. However, once the path is selected, engagement with the curriculum does not appear to vary with these individual characteristics