2022 Projects
Analysis of Safety and Security Concerns in Autonomous Vehicles
Advisor: Dr. Sen
The domain for fully Autonomous Vehicles (AVs) is a dormant area that is growing rapidly. Currently, the state of the art in AV technology incorporates the ability to sense their surroundings using hardware such as cameras, Light Detection and Ranging sensors (LIDAR), Central Electronic Component Units (ECU), and Global Positioning Systems (GPS) to perform various required functions like accelerating, decelerating, braking, turning, stopping and changing lanes to navigate roads safely. These vehicles rely on interconnected networks to be able to communicate with infrastructure, nearby vehicles, traffic lights, and road signs. As such, for a fully autonomous vehicle, the presence of a driver might be unnecessary during operation in the near future. Although such technological advancements may help reduce accidents and improve traffic conditions, from a user-centric perspective it also raises concerns about safety and security.
To embrace the fully AV technology and address the user-centric skepticism, users should be able to comprehend and trust the decisions made by the intelligent systems within the AV. To address this challenge, the objectives of this project are three-fold. First, to understand the users' concerns about the safety and security of AVs through a comprehensive social media analysis using platforms such as Twitter and Reddit. Second, to analyze the publicly available policy documents released by governing bodies such as the US Department of Transportation and the European Commission to assess the focus of the governing policies in this sector. Third, to compare and contrast the output of the first two objectives to identify the gaps and overlaps between the current government regulations and various users’ concerns and explain how these concerns are being acknowledged or need to be addressed with the use of software or hardware such that it will harbor public trust in adopting the technology of fully autonomous vehicles.
Lecture Notes |
Anomaly Detection on Smart Airport Edge Platforms
Advisor: Dr. Sen
Edge devices are decentralized devices that require real time processing carried out at the network interception. In recent years, more and more edge devices are being deployed on a network edge. With data processing done at the edge network, if there is an anomaly on the device, it affects the entire edge network. Any error existing in the outlier device could then be propagated to all connected edge devices, thereby corrupting the entire network. The aim of this project was to propose anomaly detection frameworks addressing anomalies such as API abuse, parameter tampering, and injection attacks that can be performed against APIs used on smart airport edge platforms. The frameworks address issues of latency and bandwidth that are prevalent within traditional anomaly detection without sacrificing accuracy of detection by running the anomaly detection directly on edge devices. The goal was to develop the frameworks in a generic fashion, such that they would be relevant on any edge platform, not just smart airports. This work analyzes the threats of API abuse, parameter tampering, as well as injection attacks that are prevalent in edge platform domains that make use of APIs. With these threats as our focus, we developed workflows for detecting the previously mentioned anomalies using techniques such as token buckets, black listing, hashing, and analyzing logs. Future work will include implementing our workflows in a simulation as well as potentially adding artificial intelligence or deep learning to detect anomalies within edge platforms.
Lecture Notes |
Prototype Implementation of a Blockchain-based Timesheet
Advisor: Dr. Fu
This project explores how blockchain technology can be used to enhance a timesheet system. It focuses on how a timesheet could be made more secure and tamperresistant if a blockchain is used to store the information instead of a centralized server. Many people have reported suspicion that their employer has secretly changed information on a timesheet in order to pay the employee less. The decentralized and security properties of a blockchain allow for the timesheet program to require zero trust from the employee toward their employer. Instead, all information is public and immutable, and as such, all changes are recorded publicly preventing a non-legitimate change from going undetected. This project includes a prototype software built from the Truffle suite which is a set of tools to make developing smart contracts easier to show how the concept could be implemented quickly and easily. Truffle is a command line away from being installed so that you have the required files to migrate your smart contract to the blockchain. Not only that but it allows us to easily test our smart contract using the Ganache Ethereum virtual machine. We also greatly used and edited their pet shop tutorial which has even more required files for calling Web3 to validate our transactions and sending that to the blockchain with its Javascript and Solidity files. All of this allowed us to test how well blockchains could be used for an everyday task like clocking in and out for work. The main issue we found was that current blockchain implementations are too slow and have a high gas fee per transaction. Leading us to believe this would need to be a blockchain specifically used for a timesheet across individuals.
Lecture Notes |
Risk Management in Financial Institutions: Asset Vulnerabilities & Countermeasures
Advisor: Dr. Fu
Financial institutions’ systems host a multitude of assets that need to be protected, these institutions must mitigate the threats to their assets as much as possible while also having a plan in place if their security is compromised. We identify assets in financial institutions and categorize them into the various layers of data, for example: financial, physical, operating, fixed, etc. The process of identifying, assessing, and controlling these threats to assets is called risk management. The financial institution’s risk management system comprises its policies, processes, personnel, and control systems. It is imperative that all financial institutions incorporate a sound risk management system and plan in case of a security breach. A sound risk management system identifies measures, monitors, and controls risks. Since market conditions and company structures vary, no single risk management system works for all banks. The sophistication of the risk management system should be commensurate with the financial institution’s size, complexity, and risk profile. In this project, we focus on identifying the already existing assets in financial systems and map the vulnerabilities and threats associated with financial institutions, their intellectual property, and their customers. Finally, we provide the corresponding controls and countermeasures by applying cybersecurity defense principles in depth.
Lecture Notes |