Ethan Fralick, Ryen LeBlanc, Colton McCann, Jacob Southall
Dr. Hamidreza Mirzaei & Dr. Matthew Hartmann
Streetlights are very important for community safety, but traditional streetlight
systems continue to operate with high energy consumption and other inefficiencies.
These inefficiencies lead to high energy costs and high levels of light pollution.
To solve this problem, the ProxiLux Lighting system uses a Raspberry Pi microcontroller
to manage a 100 W LED fixture with the use of Python-based logic. To manage this fixture,
the system uses a PIR motion sensor and a BH1750 ambient light sensor to detect a
pedestrian or vehicle approaching the light. Once detected, the streetlight will brighten
to 100% from its dimmed state. The ProxiLux system will keep the light fixture in
a dimmed state to reduce or decrease the problem associated with traditional streetlights.
Laboratory testing proved that the system can successfully control the light fixture
while light and motion are detected. Furthermore, a safety protocol will be implemented
into the ProxiLux system in case of subsystem failure. Ultimately, the ProxiLux Lighting
System provides a cost-effective alternative for smaller municipalities to reduce
energy consumption and light pollution.
William Fielder, Jake Pieri, Bryce Powell, Tyler Skeldon
Dr. Jinyuan Chen
Traffic congestion at intersections remains a significant problem in modern transportation systems. Most current day traffic management systems rely on preprogrammed fixed-time control methods that fail to adapt to fluctuations in traffic conditions, which often leads to inefficient traffic flow and increases in travel time, fuel consumption, and unnecessary vehicle emissions. A dynamic traffic control system offers a solution to the traditional fixed-time system’s limitations by introducing machine learning to optimize signal timing based on real-time traffic data. The proposed adaptive system utilizes cameras to detect traffic conditions which are used as input data for a microcontroller running a machine learning based algorithm that dynamically adjusts traffic signal phases to improve intersection efficiency. System performance was tested and analyzed using both a traffic simulation environment and a small-scale physical prototype. Results from testing this system demonstrate that adaptive traffic control can reduce average vehicle waiting times and improve traffic efficiency when compared to traditional fixed-time systems. The research and results of this project contribute to ongoing advancements in artificial intelligence and its future applications in intelligent infrastructure.
Preston Holwager, Yahir Levario, Matthew Maggio, Alan McGhee and Kelton Sepulvado
Dr. Hamidreza Mirzaei
Efficient monitoring of soil conditions is essential for improving crop health, efficient irrigation, and productivity. The Soil & Crop Monitoring System is built to allow farmers to have access to continuous real-time readings of their current soil conditions, allowing them to make informed decisions. Our device measures soil moisture and pH using appropriate sensors, which then sends the information to an ESP32 microcontroller. The system then transmits this data to a Raspberry Pi where the data is displayed and stored in a CSV file. This data can then be implemented into a MultiVariate Time Series Forecasting with LSTMs that helps predict irrigation needs based on current soil conditions. The current readings and predictions can be accessed by a dashboard powered by a Raspberry Pi.
Bryant Moore, Gabriel McMillian, Jason Jones, Patrick Day
Strikewerx
Hannah LeBlanc and Jason Wagner
Dr. Matthew Hartmann
Air Force Global Strike Command personnel operating in hazardous environments require
restrictive Mission Oriented Protective Posture gear, including the M50 gas mask.
Consulting printed technical manuals while wearing thick butyl rubber gloves creates
operational hazards, forcing operators to break visual contact with their task and
risk cross-contamination. To resolve this, a wearable Heads-Up Display (HUD) was engineered
directly into the M50 mask's visor. The offline system integrates Commercial-Off-The-Shelf
RayNeo birdbath optics driven by a Raspberry Pi Zero 2W and a Geekworm power management
system housed in a belt holster with a tactile, glove-friendly four-button interface.
Initial prototype testing validated the microcomputer's capability to render complex
PDF manuals at 30 FPS without memory overflow. Power analysis shows a battery life
that achieves over 8 hours of usage. This self-contained HUD proves the feasibility
of hands-free data visualization in contaminated zones without compromising the mask's
protective seal.
Christian Buck, Slade Foil, Laterrence Harris, & Bryce Trapen
Dr. Jinyuan Chen
Terahertz communication is an advancing technology for the next generation of wireless networks: 6G. It offers larger bandwidths capable of producing ultra-high data rates, much more so than previous generations. High frequency THz communication systems experience more non-ideal properties that affect signal propagation, requiring much larger numbers of transmitting antennas and highly precise directional beamforming to establish and maintain connections. This makes THz systems computationally complex, making traditional beamforming techniques inadequate. This project addresses the optimization needs of THz communication by developing a Python-based simulation environment employing cutting edge, deep learning techniques that drastically increase the functionality of beamforming in THz systems. The result of our trained model outperforms modern beamforming techniques regarding signal to noise ratio and throughput, correlating to higher accuracy and higher throughput, respectively. The project’s findings verify and increase the effectiveness of deep learning implementation for future 6G wireless network designs, supporting sustainable and intelligent communication infrastructures.
Carter Boone, Holden Clark, Henlee Hoffer, Landon Morreale
Dr. Matthew Hartmann and Dr. Arif Hussain
Electrical engineers are tasked with allocating power to meet the demands of consumers’ homes, which produces a monthly utility bill. Our goal is to decrease the cost of that bill by eliminating standby power, which is caused by appliances that are plugged in and not being used drawing small amounts of current. To achieve this, our team developed a Smart Home Energy Management System (SHEMS). This system is a network of hubs that tracks devices plugged into our simulated home, with the capability of remotely shutting them off when not in use. Other SHEMS on the market are usually expensive and lack system autonomy. During our engineering process, we wanted to combat these things by using more cost effective components, implementing machine learning, and allowing the user to take full control of their energy usage. We found that our device works similarly to others on the market in its ability to track crucial data regarding power consumption while being more cost-effective.
Isaiah Sharplin, Willie Verret, Caleb Worsham and Cameron Westerfield
Dr. Lingxiao Wang and Dr. Jinyuan Chen
The goal of this project is to improve the quality of life for patients and medical personnel by introducing a dynamic stretcher system to mitigate the downsides of conventional stretchers, namely poor shock absorption and limited mobility over inclines. With the “All Terrain Gurney", we hope to improve traditional stretcher design by implementing real-time adjustment via motorization and software integration. This would enable safe and comfortable transportation for patients up to 250 lbs. The core of our project is the implementation of a PID controller system which measures feedback from multiple accelerometers with an established x and y setpoint. The accelerometers detect acceleration changes in the x and y directions, which are sent to the PID to perform live calculations. These calculations are then sent to a pair of motors, so that necessary adjustments can be made. This system is not limited to the application we chose. It can be applied to any application requiring dynamic platform adjustment, particularly when stable transportation is necessary.
Andrue McMillan, Julian “Karo” Oghomi, Gray Sharp and Cameron Ytzen
Dr. Matthew Hartmann
The Eco-Electric Bike Rack is an efficient, eco-friendly solar powered charging station for all types of electric bikes. This project intends to address the rising usage of electric bikes across campus and their lack of available public charging stations in an eco-friendly way. In order to achieve this goal, solar panels efficiently capture the sun's energy using single-axis solar tracking technology and store it into a DC battery. This battery then provides power to a charge controller and two custom-built variable boost converters to allow for the charging of up to two bikes of varying voltages simultaneously. With these systems, this charging station allows for virtually every type of electric bike to be charged up to the rated manufacturers speeds with up to two bikes charging simultaneously. The development of the Eco-Electric Bike Rack will help alleviate the lack of charging stations and address the rise of electric bikes. Furthermore, the design of this charging station could potentially be upscaled to allow for its implementation into larger urban and suburban areas.
Ethan Ice, Michael Macaluso, William Martin and Troy Mosley
Parish Controls
Blaine Russ, CEO
Dr. Miguel Gates, Dr. Jinyuan Chen
Our project seeks to advance the way PLC panels are fabricated and reduce the cost and stress of building said panels. PLCs are used to automate and control vast systems and devices from gas plants to paper mills. All current devices used in the making of PLC panels are very basic in their operation, but our device has taken the base idea of these basic devices and has enhanced it with the use of a centralized information hub via a mobile touchscreen which is connected to a Arduino GIGA micro controller. With the touchscreen and the micro controller connected to our rotary Hall Effect sensors used to measure wire length used in the fabrication of the PLC panels, the heavy duty linear actuator for lifting the PLC panels for ease of wiring, and the Hall Effect limit switches used in the detection of the rack's position, we are able to control our entire system with the touch of a button. Combining all of these subsystems onto our rack fabricator while keeping the rack mobile via bottom wheels and powered simply from a wall outlet, we are able to greatly increase the speed, safety, and lower the cost of making PLC panels.
Amir Cazabat, Ian Dryg, Hojun Lee, Christine Meister
LaSPACE
Dr. Sandra Zivanovic and Dr. Matthew Hartmann
Memristors, or memory resistors, are electronic devices that change resistance based on current flow. They are gaining attention for their simple 2-terminal structure, low power consumption, and neuromorphic computing capabilities. Memristors address scalability and density challenges of conventional CMOS-based memory and computing by using memristor arrays in resistive random-access memory (ReRAM). This project aims to enhance cobalt ferrite (CoFe2O4) memristor arrays to store more data in a smaller form factor. Previous CoFe2O4 memristor arrays had issues with inconsistent resistive states and crosstalk between devices. To address this, sputtering was used instead of the sol-gel process to deposit thin films and various array structures were designed and tested to minimize crosstalk. We anticipate achieving more consistent resistive states and minimal to no crosstalk without compromising other parameters like resistance ratio due to differences in fabrication methods and array design.
Owen Brinson, Walker Badon, Elijah Hill
Dr. Jinyuan Chen
New generations of cellular technology generally emerge about every 10 years with 4G cellular becoming commercially available around 2009, and 5G rolling out around 2019. With 6G technology available a few years in the future, it has some significant advantages over the previous generations however it is not without its limitations. One of the main issues with 6G is that because it will operate in the terahertz frequency range, the wavelengths will be extremely small, and the signal will be more easily blocked by physical obstacles. Our project aims to enhance 6G wireless communications by fixing this issue. We plan to use a physical passive reconfigurable intelligent surface (RIS) to demonstrate its effectiveness in boosting signal strength in occluded environments. We will also simulate an RIS-assisted wireless link using machine learning in Python to optimize signal propagation and spectral efficiency. We will compare the results of our machine learning model to other machine learning models as well as a baseline model to show performance increase. These results will show that 6G communications can be further improved by implementing machine learning into an RIS-assisted system with minimal long-term human intervention.
David Allen, Caden Britt, Toby Latino, Conner Saucer
Jeremy Blair, SEL Application Engineer
Dr. Prashanna Bhattarai and Dr. Matthew Hartmann
Our signal generator is a testing device specifically designed for the SEL-734B capacitor bank controller. The SEL-734B controller is used by utility companies and industrial sites to switch capacitor banks onto distribution lines for power factor correction. This specific controller operates based upon the measured power factor. Our signal generator serves as an alternative to existing relay test-sets, satisfying the need for a portable and cost-effective piece of test equipment. Our approach involves generating a signal to replicate the current through a distribution line, adjusting its phase shift relative to the system reference, adjusting its magnitude, and passing it to a 734B. We also pass a voltage signal directly from the grid to the 734B. Using the output signals generated by our device, the end user can replicate the varying conditions of a distribution level power system. Our signal generator provides an output range of 0 – 15Vac, any desired phase shift, and a meter socket adapter for use in the field. We believe our device can be used to minimize the time, cost, and complexity of capacitor bank controller commissioning.