This project has developed a testbed for evaluating novel algorithms for spectrum sensing in integrated sensing and communication applications. Primary and secondary users must be able to share channels and switch between them as necessary. Algorithms facilitating these processes are being researched and/or developed. However, a generic testbed currently doesn’t exist where these algorithms can be tested against simulated signals and this project aims to remedy this issue.

Team Members

Generative Adversarial Networks (GANs) are a class of machine learning frameworks that automatically studies the regularities in the input data and uses the pattern to generate new datasets from the original one. They are composed of two parts: a generator model and a discriminator model. The purpose of a generator model is to generate data, while the purpose of a discriminator model is to distinguish between the generated and real data. An adversarial network is formed by connecting the discriminator outputs to the generator, allowing the models to compete.

Design and implementation of a Physical Unclonable Function (PUF)-Based Perceptual Image Hashing Algorithm used to verify the authenticity of an image and its device of origin.

Team Members

Abdullah Haris, Dennis Liu, Jason Math, Brogan Neagle, Paul Thompson

A pilot’s vigilance level can be assessed using EEG (electroencephalogram, measures brain signals) and EOG (electrooculography, measures eye signals), which are collected via physical electrodes placed around the user’s head. Based on the U.S. Army’s use-case of collecting data mid-flight, our EEG/EOG measurement system must be lightweight, low-profile, and accurate. However, accurate EEG/EOG headsets (used in academic environments) are large and obtrusive, while sleek EEG/EOG devices (sold commercially) are inaccurate.

In this project, we focus on training and compressing a neural network to implement reliable channel codes for use in both simulated and real world communication scenarios. The project started with a large neural network that was developed by prof. Hyeji Kim and then compressed the neural network by changing hyperparameters, implementing distillation, and using single value decomposition to produce smaller and more efficient algorithms.

Team Members:

Nicky Dahl, Mathew Puente, Sam Rizzo, Ryan Root, Allen Shufer, Jonas Traweek

Recent advancements in machine learning and computer vision have created an environment in which new applications in accessibility design have arisen. To fully utilize this technology, we worked with the Laboratory for Image and Video Engineering (LIVE) at the University of Texas at Austin to create a React Native based mobile application that will assist the visually impaired in taking better photos for social media, which has become a predominant aspect of daily life for a large percentage of the world.

We have developed an application to give people a better way to judge the value of their

The Identity Wallet is a web application that makes it possible for individuals to keep track of and manage their identity credentials in a specialized digital wallet. Based on the UT CID Identity Ecosystem, the Identity Wallet generates a personalized list of personal data on account creation and allows management of that data, as well as logging of companies an individual has shared specific data points with. The Identity Wallet also provides risk and liability calculations and sharing recommendations regarding specific personal data using probabilities based on real-world events.

pTOLEMY provides an innovative solution for the monitoring of large industrial pipe networks. As the device flows through the pipelines, it will collect positional data. Once the device is retrieved from the pipe network, the user can use our desktop app to easily analyze the data and see if there are any issues within the pipe network. pTOLEMY can help catch issues within the pipe network before they become potentially costly and dangerous.

Team Members

Chloé Arana, Arshad Bacchus, Sachin Desai, Indhar Gopalakrishnan, Kevin Mechler, Andy Ni

Many students don’t feel safe when walking alone on college campuses and have had encounters that make them feel uncomfortable. EmBrace is here to provide confidence and protection, and for personal safety against threats. It is a safety device that incorporates a wearable device and an app. EmBrace will vibrate to alert you that someone is approaching fast. Based on the device’s warnings, the user can press the stop button on the physical device to dismiss the vibration or send an emergency text using the emBRACE app.