41530 research outputs found
Sort by
Sentry V3: Extending Context Switches on a Trusted Secure Coprocessor
Get PDFSoftware correctness and integrity is only ensured through trust in the un- derlying hardware. However, modern computer systems are complex to design and secure. Thus, given the choice between performance and security, companies will often prioritize performance, resulting in vulnerable systems. This creates exploitable systems that must be patched retroactively because business value performance over security. One approach to this issue is to separate the root of security from the rest of the system to create a minimal trusted computing base. Trustguard is one instance of this. Trustguard implements a Containment Architecture with Ver- ified Output (CAVO) model which shows how a simple, pluggable co-processor, called the Sentry, can secure commodity systems. The Sentry monitors com- mitted instructions from trusted software to enforce containment and software integrity. The Sentry is currently implemented as a proof-of-concept single-context co- processor for unicore systems with no concurrency features. This thesis aims to extend the work of the Sentry by leveraging its correctness guarantees and present a design capable of managing multiple single-threaded programs. It discusses the problems and solutions when managing multiple independent Merkle trees to ensure program integrity and isolation for processes on a time-sharing processor. To this end, this thesis proposes a new Sentry hardware architecture, runtime algorithms, and bootstrapping procedures to support context switching
Data-Driven Product Recommendations: A Decision Support Framework Utilizing Customer Reviews
Get PDFWith the considerable presence of e-commerce in society, vast number of purchasable goods, and increasing brand variety, consumers are faced with the challenge of buying products that they perceive to be of greatest value to them. To assist consumers with making better informed decisions, e-commerce websites allow individuals to post their own experiences and score the products that they purchase. Despite this information, the variety of experiences and feedback that consumers share do not always lead to clarity on whether a product is best suited for the purchaser. To help guide customers through a simplified purchasing process from the perspective of an online retailer, this study seeks to explore a possible framework for combining existing quantitative product metrics and qualitative reviews with a recommender system. Results are then compared with recommendations provided by the well-known ChatGPT language model, which is also emerging as a trusted decision-making system. 10 different product categories, each containing 8 to 10 products, had their customer review data aggregated and analyzed for quantitative sentiment scores. The resulting data was mixed with common numerical product attributes and analyzed with common multi-criteria decision making models to form organized product recommendations. The final recommendations demonstrated some discrepancies from ChatGPT. However, our framework demonstrated that a sentiment-inclusive recommender system can be established with similar performance to complex models such as ChatGPT, and that further experimentation on feature selection could improve recommendation performance
Efficient Gan-Based Adversarial Example Generation Against Ml-Based Network Intrusion Detection Systems
Get PDFIn the realm of network security, Network Intrusion Detection Systems (NIDS) are essential for identifying and mitigating malicious activities targeting networked devices. Traditionally, these systems have relied on signature-based and anomaly-based detection techniques. However, the increasing complexity and adapt- ability of cyber threats have driven the adoption of Machine Learning (ML) ap- proaches in modern NIDS, significantly improving their ability to detect a wider range of attack vectors. Despite these advancements, ML-based NIDS remain vulnerable to adversarial examples—deliberately crafted inputs designed to mislead models and trigger incorrect classifications. Originally identified in the field of computer vision, adversarial examples now pose a critical threat to the reliability and robustness of ML- based cybersecurity systems. This thesis explores the use of Generative Adversarial Networks (GANs) to generate adversarial examples that closely resemble legitimate network traffic while effectively evading detection. By leveraging the generative power of GANs, the research aims to produce realistic and functional adversarial samples efficiently. The primary objective is to adapt, enhance, and evaluate a proposed GAN-based adversarial example generation algorithm to improve its effectiveness in testing and validating the resilience of ML-based NIDS. Experimental tests demon- strate that the proposed algorithm outperforms existing methods, achieving better results with reduced computational overhead
Poly Mars Mini-Rover
Get PDFThe Cal Poly Mars Rover project is creating a fleet of near-identical rovers operating collaboratively to research the surface of Mars. Over the last six years, teams at Cal Poly have been developing vehicles to be launched upon the commercialization of interplanetary travel. Two of these vehicles are to take the roles of ‘Scout’ and ‘Relay’ with the most significant difference being their respective ‘arms.’ The ‘Scout’ will have a robotic arm that interacts with objects on the surface while the ‘Relay’ will have a fixed arm, equipped with a StarLink antenna to communicate with orbiting satellites as seen in Figure 2. The primary goal of this project is to redesign the rovers’ arms and arm deployment system, and to design the attachment mechanism for the antenna. The Poly Rover lead, Professor Rich Murray, acts as the primary stakeholder for the Senior Project Group listed above
2-Axis Solar Panel Tracker
Get PDFSolar panels generate energy from sunlight. To do this it’s important that they remain pointed at the sun throughout the day. The Cal Poly Micro Grid Lab runs a class that uses a student made solar panel and tracking mount as an instructional tool. This project is beginning to show signs of wear from outdoor use and is unable to run without constant input due to some significant design flaws. The lab needs a new version of the mount to be made that is significantly more reliable and easier to use
Heart Valve Leaflet Cutting Automation
Get PDFThe content contained in this report is confidential and proprietary information of Edwards Lifesciences
Project Owl: QuAD RC2
Get PDFQuad RC2 is a new radio designed to be compatible with the Cluster Duck Protocol (CDP). It must be able to act as all parts of the mesh network and be able to find its location with GPS. Much of the RC2 effort is already completed. This project ports the previously developed firmware to the new board and runs tests to assess its capabilities
Expeditionary Ocean Power Generator
Get PDFPoint Absorber Wave Energy Converters are currently being explored as a new form of energy harvesting. The Naval Facilities\u27 Expeditionary Ocean Warfare Center (NAVFAC EXWC) has commissioned an interdisciplinary project for students at California Polytechnic State University, San Luis Obispo, to design a prototype Expeditionary Ocean Power Generator (ExOPG) for charging batteries in the ocean.
The Mechanical Engineering team is responsible for assembling the prototype chassis, which converts the vertical bobbing motion of a buoy into kinetic rotational energy that can be used to rotate a generator shaft. They are additionally tasked with isolating the electrical components from the harsh ocean environment. The Electrical Engineering team is responsible for designing a battery charging system that takes rotational energy from the mechanical subsystem and safely and efficiently converts it to electrical potential energy, which is then used to charge an onboard payload battery.
The system was tested in a near-shore environment in Morro Bay, California. While each system functioned successfully independently during bench tests, and all components from each team survived the harsh saltwater environment, the prototype system did not significantly charge the battery in the ocean. Key factors were identified that caused the system to fail both teams, and with minimal design changes, the entire system can be tested in the future with more promising results
Electrocardiogram Signal Modeling and Classification Using Machine Learning
No full textThis project studies electrocardiogram (ECG) signal classification using various machine learning technologies, including Random Forest (RF), K-Nearest Neighbor (KNN), Support Vector Machines (SVM), and Convolutional Neural Networks (CNN). Computer simulation results show these algorithms perform well on the MIT-BIH database, with accuracy above 98%
Effect of Manufacturing and Machining Process on Recycled Composite Connecting Rods Under Tension and Compression
Get PDFCarbon fiber reinforced polymer (CFRP) is an engineering material renowned for its high-strength and low-density material properties. Such properties make CFRPs especially useful for applications across many industries, including aerospace, automotive, marine, sporting goods, oil and gas, and more. However, with such favorable material properties comes the cost of high energy usage required to produce raw carbon fiber. Recycled carbon fiber (rCF) has emerged as a promising alternative, offering a significant reduction in energy usage. This research explores the viability of applying recycled CFRPs to internal combustion engines to both reduce energy costs during manufacturing and while in use. Mechanical properties of two CFRP materials were characterized through tension and compression testing. Composite connecting rod samples were fabricated utilizing autoclave curing and precision CNC machining, then tested in representative load cases simulating engine operation. Finite Element Analysis (FEA) models were developed in Abaqus to simulate and predict the mechanical behavior of the connecting rods and validated against experimental results. Results of this study showed that CFRPs exhibit mechanical performance similar to aluminum. While woven prepreg CFRPs achieved the highest performance, rCF demonstrated sufficient structural qualities to be considered a viable, sustainable engineering material. This work highlights the potential of recycled composites to reduce manufacturing emissions and support sustainable engineering practices with composite materials