ECE Undergraduate Laboratory
Senior Design Showcase
Welcome to the ECE Senior Design Showcase, where innovation meets real-world problem-solving! This event highlights the creativity, technical skill, and dedication of our students as they present their final projects to a panel of judges.
From AI-powered security solutions to cutting-edge robotics, sustainable energy systems, and even brain-computer interfaces that explore thought classification, these projects tackle challenges that impact industries and everyday life. Each team has worked tirelessly to bring their ideas to life—designing, building, and refining solutions that push the boundaries of engineering.
Explore the diverse range of projects presented this semester.
The Projects
Team
Voice-based network configuration, utilizing Large Language Models (LLMs), is emerging as a valuable approach for automating and streamlining network management. KIRA serves as a platform that streamlines Cisco device configuration operations by handling the difficulties faced during Cisco IOS usage. Voice commands entered by the user generate scripts that the system executes securely through SSH connections toward network equipment.
This approach significantly reduces the time, cost, and potential for human error in resolving common networking issues, making it particularly valuable for small businesses and educational institutions that may lack full-time network specialists.
Advisor: Dr. Roberto Rojas-Cessa
Team
Potholes are a significant issue for road maintenance, causing safety hazards, vehicle damage, and increased repair costs. Current methods of pothole repair and road maintenance primarily rely on reporting from drivers and the dispatch of road maintenance workers to assess and repair the pothole.
We aim to automate pothole reporting and assessment in a system that detects potholes using a real-time object detection algorithm YOLOv8 and obtains a rendering of the pothole using Light Detection and Ranging (LiDAR) sensor which scans the road from a moving vehicle. When the pothole is detected, it is mapped to a Global Positioning System (GPS) location, and measurements of these potholes are analyzed to determine the severity and potential repair cost for each one. All of this data is stored in a database and synthesized into a digestible web interface for government authorities to act upon.
This project would benefit local governments, road maintenance services, and private contractors seeking automated solutions for road inspection and determining conditions for repair.
Advisor: Dr. Xuan Liu
Team
A sonic fire extinguisher is a device which extinguishes fires using low frequency sound waves. The waves, causing acoustic pressure and air velocity, displace the oxygen molecules that fuel the fire, thus extinguishing it. The sonic fire extinguisher is built using a subwoofer (for low frequency sound), an amplifier, a power supply, and a collimator, which focuses the sound waves.
This fire extinguisher provides an alternative to traditional fire extinguishing methods, such as those utilizing chemicals or water. Therefore, it has a practical application in electrical fires, and critical environments such as semiconductor fabrication and data centers.
Advisor: Dr. Xuan Liu
Team
As electric vehicles become more common, the need for sustainable charging stations in areas without access to the power grid continues to grow. To address this, we designed a fully off-grid EV charging station that uses solar power, battery storage, and Level 1 (120V, 16A) SAE J1772 car charging. It features six 200W solar panels, generating an average output of 6kWh per day with around 5 hours of direct sunlight. These are connected through an MPPT-enabled inverter and charge controller to charge a 12V battery bank. An Arduino-based control circuit handles communication with the vehicle by generating the J1772-compliant pilot signal and switching the relay based on detected charging states.
The project demonstrates a fully functional prototype of solar EV charging and lays the groundwork for potential Level 2 expansion and smart grid integration.
Advisor: Dr. Xuan Liu, Marcos Neto
Team
HomeGuard is a smart home automation system designed to optimize comfort, energy efficiency, and security through intelligent automation. Our solution integrates multiple sensors—including temperature, humidity, light, and motion—with an ESP32 microcontroller and AI-based logic to deliver seamless control over essential home functions. Users can interact with the system via a smartphone interface, allowing for both manual control and real-time monitoring. Unlike commercial alternatives, HomeGuard emphasizes local data processing, hardware customization, and full data privacy—offering custom automations without vendor lock-in or monthly fees. Through machine learning, the system adapts to user habits over time, predicting and automating tasks such as lighting, temperature control, and security alerts.
HomeGuard is scalable, low-cost, and designed to make smart living accessible, especially for individuals with mobility or accessibility needs.
Advisor: Dr. Xuan Liu
Team
The project, Earthernet: Large-Language Model-Aided Smart Agricultural Care System with Internet of Things Applications, aims to develop a smart agriculture system utilizing real-time data from sensor modules and relevant context retrieval. Through the use of solar energy and housed in a 3D-printed enclosure, the system consists of a main Raspberry Pi that will communicate wirelessly with an ESP32 module to monitor key environmental factors such as soil pH, humidity, and specific nutrients using a specialized sensor. By leveraging fine-tuned large-language models with retrieval augmented generation capabilities, the project will provide users with accurate, generated responses into the plant’s condition, potential care recommendations, and other useful information. A dashboard displays graphs and tabular data for the user to further examine the figures.
This system aims to optimize plant care by offering easily understandable and tailored suggestions based on collected data, enabling more efficient and informed practices for backyard gardeners.
Advisor: Dr. Cong Wang
Team
The goal of our project was to design and build an AI-powered fingerprint recognition security door that enhances security, accessibility, and smart home integration. By combining a biometric fingerprint sensor with a motorized electric lock and the NVIDIA Jetson Nano, our system enables real-time, keyless authentication with high accuracy. Machine learning is utilized to improve recognition reliability, meeting industry standards for both False Acceptance and Rejection Rates. Additional features such as a speaker for audio feedback and a backup power supply further improve the system's functionality and resilience. This solution is ideal for residential and small commercial applications. It offers a modern, secure, and user-friendly alternative to traditional locking systems.
Advisor: Dr. Xuan Liu
Team
In this project, we designed a wireless, motion-controlled, robotic arm. Using object recognition technology running on a host device, the arm can be controlled from a distance via user hand movements. We used high-definition cameras, a Raspberry Pi Model B, servo motors, and a 3D printed robotic arm with six degrees of freedom (DOF) movement capabilities to complete the project. The arm, which has a custom (non-humanoid) gripper, is intended to mimic the functionality of industry standard robotic arms for jobs that require high dexterity and work with small areas of interaction, such as surgery, non-automated industrial assembly, and bomb defusal.
Our robotic arm aims to fill a unique niche in the market by offering advanced functionality and wireless control at an affordable price, ideal for small businesses as well as educational, healthcare, industrial facilities in developing nations. By being under 7 pounds in weight, having a stable power consumption of less than 10 watts (5V, < 2A max) and being offered at a sub-$200 price point, our design provides comparable functionality and precision without excessive costs and logistical demands.
Advisor: Dr. Xuan Liu
Team
Our senior project focuses on developing a thermal wind tunnel designed to test electronic components under a range of wind and temperature conditions, simulating environments such as server rooms or desktop systems. The tunnel will be self-regulating and user-friendly, allowing users to input desired temperature and wind speed values, which the system will automatically maintain. The design incorporates five key features: temperature control, fan speed control, environmental sensing and monitoring, wind speed regulation, and integrated automated control. The thermal system will include a safely integrated heating element, while the airflow will be managed using a fan controlled by a stepper motor or variac. Motorized dampers will adjust airflow, and an anemometer will monitor wind speed, providing feedback for automatic regulation. All systems will be controlled by a microcontroller or Arduino to ensure synchronized automation. This project is inspired by industry interest in testing voltage regulation technologies under variable thermal and airflow conditions.
Advisor: Dr. Dong-Kyun Ko
Team
For our project, we propose a mobile battery-operated vehicle with a sensing and tracking system to actively detect the ratio of parked cars in a parking garage. Our project improves upon existing systems by being cost-effective while having the ability to check every parking spot rather than only monitoring the access points. This system will utilize cameras and machine vision using neural inference to detect vehicles. Our platform will be constrained to an overhead track and have a docking station.
Besides providing residents with a detailed and real-time view of available spots, this system can be an important analytics tool for managers and administrators.
Advisor: Dr. Xuan Liu
Team
This project is focused on a human-following autonomous robot that will help transport excess loads for the ease of the user. The designated purpose is to serve as an everyday assistive tool to the demographic audience of people with mobility issues and parents with young children. This robot is constructed with the Arduino Uno and ESP-32 with camera microcontrollers and multiple sensors coded for obstacle detection and avoidance. Its construction has an easily storable chassis with a removable cargo hold. The working principle is to have the microcontrollers work in tandem with the sensors to have the autonomous robot follow the human by pinpointing the location and direction of a particular shape and color.
Advisor: Dr. Xuan Liu
Team
In this project, we aim to create a wireless piano LED visualizer by establishing communication between a musical instrument digital interface (MIDI) and a microcontroller. Our goal is to assist teachers in both in-person and remote piano lessons, helping beginner students recognize and visualize notes directly on the piano. This tool is intended to attract educators and learners who benefit from visual, interactive, and remote learning approaches.
Advisor: Dr. Xuan Liu
Team
This project involves the development of a low-cost, portable Brain-Computer Interface (BCI) system capable of classifying directional thoughts using EEG signals. The system uses wet electrodes to collect neural activity, which is then filtered, amplified, and processed using a neural network deployed on embedded hardware. Emphasis was placed on creating a fully battery-powered solution suitable for assistive technologies and hands-free control in hazardous environments.
The project addressed challenges such as signal noise, hardware integration, and real-time classification within strict economic and technical constraints. This work not only advances accessible BCI development but also lays the groundwork for future expansions including wireless communication, increased classification accuracy with more electrodes, and potential patentable innovations in affordable neurotechnology.
Advisor: Dr. Dong-Kyun Ko
Team
This project presents a low-cost, embedded Driver Drowsiness Detection system using computer vision on a Raspberry Pi 5. A USB camera captures real-time facial input, and early signs of fatigue are detected using Eye Aspect Ratio (EAR) and Mouth Aspect Ratio (MAR) derived from 68-point facial landmarks. OpenCV handles video processing, dlib performs landmark detection, and SciPy is used for geometric calculations. Visual alerts are shown through a traffic light-style LED module, while audio warnings are played via a speaker. The system is fully autonomous, launching on boot and operating without external hardware.
Challenges addressed include maintaining low latency, ensuring accuracy in varying lighting, and achieving stable peripheral initialization. The final prototype operates independently of external computing resources and is suitable for applications in automotive safety, educational demonstration, and embedded vision prototyping. This work demonstrates a scalable foundation for real-time, affordable driver monitoring solutions in resource-constrained environments.
Advisor: Dr. Dong-Kyun Ko
Team
Electric guitarists often face limitations when attempting to integrate their instruments with analog modular synthesizer systems. Most existing solutions are monophonic, meaning they are unable to handle multiple notes simultaneously, restricting guitarists from playing chords. Our project addresses this limitation by introducing an interface that processes each string of an electric guitar independently in real-time.
This is achieved using an STM32 microprocessor that performs six Fast Fourier Transforms (FFTs)—one for each guitar string—to extract pitch and volume data. That data is then converted into control voltage (CV) signals, a standard format for representing musical information within analog modular synthesizer systems. Users can route these CVs to a variety of modules such as voltage-controlled oscillators (VCOs), voltage-controlled amplifiers (VCAs) and voltage-controlled filters (VCFs) to create complex and unique sound designs. By offering low-latency, polyphonic guitar-to-CV translation in a compact embedded platform, this device allows guitarists to interact with modular synthesizers in a way that previously required keyboard or sequencer-based systems.
The device supports market expansion by enabling direct engagement from the broader population of electric guitar players, while remaining fully compatible with the modular synthesizer ecosystem it is intended to be sold into.
Advisor: Dr. Xuan Liu
Team
In this project, we made a parking assistance system using ultrasonic sensors that inform the driver of nearby objects. We used sensors to detect the distance to the nearest object from the bumpers, and send that information to a microprocessor to display the information using visuals and audio. The system also contains an emergency automatic braking piston that presses the brake pedal before an imminent collision. The design includes ultrasonic sensors, a microprocessor to handle data, an LED display to provide visuals, speakers for audio cues, and a piston for pushing the brake.
This could be sold to the open market as a universal implementation or to companies that need cheaper designs for affordable cars.
Advisor: Dr. Dong-Kyun Ko
Team
My project details the design and implementation of an In-Circuit PCB Automated Test Equipment (ATE) system intended to streamline and automate the testing of printed circuit boards (PCBs) during manufacturing and validation. The developed ATE uses a bed-of-nails style test fixture with 48 precisely aligned spring-loaded pogo pins to establish electrical contact with the Unit Under Test (UUT). An Arduino Mega 2560 microcontroller orchestrates the control of three MAX306 analog multiplexers, enabling high-speed switching across 48 discrete analog channels. Voltage measurements are acquired through a high-resolution 16-bit ADS1115 ADC, providing microvolt-level sensitivity and ensuring accurate data capture. Measurement results are displayed in real-time on a 5-inch Nextion capacitive touchscreen, with pass/fail status determined against programmable voltage upper and lower limits for each channel. The ATE system features fully automated push plate actuation using motorized linear actuators, supported by gas struts to ensure smooth, controlled movement and prevent mechanical shock to the UUT. Measurement data is logged continuously onto a 32GB microSD card for traceability and quality assurance documentation. The design integrates mechanical, electrical, and software subsystems into a compact, user-friendly platform powered by a 120V AC source with internal fusing and active cooling. By combining modular hardware with automated testing workflows, this ATE provides a scalable and affordable solution for mid-larger volume PCB manufacturers, quality assurance teams, and repair centers seeking faster, more reliable board validation.
Advisor: Dr. Xuan Liu
Team
Drones, also known as Unmanned Aerial Vehicles (UAVs), are becoming more popular in areas such as environmental monitoring, search and rescue, and infrastructure inspection. Our research aims to develop an autonomous drone optimized for energy-efficient flight, capable of water sampling and monitoring. The drone will integrate our EcoFlight optimization pathing model, designed to minimize energy expenditure in flight travel and obstacle pathing, increasing overall mission time. This type of unmanned vehicle will necessitate amphibious takeoff and landing for secure aquatic sampling. The system will support live sample collection, real-time data transmission, and the ability to support a variety of sensors to enhance sampling capabilities. Overall, this project aims to improve accessibility and efficiency of water sampling, reducing the time and labor required for sampling of hard-to-reach locations. Enabling higher precision water monitoring and easier tracking of contaminants.
Advisor: Dr. Roberto Rojas-Cessa
Judges Panel
Advisors
