This meeting will feature presentations on select senior projects by students from University of Cincinnati, Miami University, and Northern Kentucky University.  Each year the engineering students complete their senior projects.  We are amazed at the talent and creativity of these students as they present their projects to us. These presentations are typically an interesting mix of hardware and software, solving problems, improving efficiencies, and creating new opportunities.

ABOUT THE PRESENTERS:

Miami University

Project Title: Autonomous Omni-Directional Automated Guided Vehicle (AGV)

Team Members: Seth Burghard, Nick Delaet, and Mason Powers

Faculty Advisor: Dr. Mahdi Yazdanpour

Abstract: The objective of this project is to design, prototype, and demonstrate an autonomous omni-directional Automated Guided Vehicle (AGV) capable of navigating both indoor and controlled outdoor environments on a single-floor surface. The AGV will utilize a holonomic drive system using Mecanum wheels, a LiDAR sensor for Simultaneous Localization and Mapping (SLAM), and an NVIDIA Jetson Nano running the Robot Operating System (ROS) for localization, path planning, and obstacle avoidance.

 Project Title: Virtual Twinning Progression to Analyze/Predict Circuit Failure

Team Members: Logan Liu, Sam Shuman, and Sean Whyle

Faculty Advisor: Dr. Mark Scott and Dr. Peter Jamieson

Abstract: Capacitors play critical roles in power conversion applications.  Yet they are among the components most likely to fail.  This project validates a methodology created to observe failing capacitors.  It created a model using both old and new experimental data and verified its accuracy.

Project Title: Online Programmable Logic Controller Course

Team Members: Charlize Hadix, Philip Hampton, Carter Smith, and Brandon Vu

Faculty Advisor: Dr. Mark Scott and Jim Leonard

Abstract: Our project is dedicated to providing a hybrid course on Programmable Logic Controllers (PLCs) for Miami University students. The course covers various topics regarding PLCs such as its history, safety, programming, I/O, etc. to allow students to be prepared for PLC use in an industrial environment. Hands-on labs were also implemented to help students further develop the skills necessary for operation as well as gain practical experience with PLCs.

University of Cincinnati

Project Title: Ensuring UAS Airworthiness: Deep Learning Based Acoustic Health Monitoring of Motor Health

Team Members: Siddharth Urankar, Prissha Chawla

Faculty Advisor: Dr. Manish Kumar

Abstract: This work presents an inflight UAV powertrain health monitoring framework using machine learning for real time anomaly detection. We collected high fidelity acoustic signatures from Brushless DC motors to develop a semi supervised 1D Convolutional Neural Network Autoencoder. By training exclusively on healthy acoustic profiles, the system identifies mechanical degradation by analyzing reconstruction error thresholds during flight. This non invasive approach supports preflight checks and active monitoring within the In time Aviation Safety Management System (IASMS) to ensure airworthiness in mission critical environments.

Project Title: UC Navvy

Team Members: Elaine Mansour, Justin Lin, Karr Stump

Faculty Advisor: Giovani Abuaitah

Abstract: UC-Navvy is an indoor and outdoor campus navigation system designed specifically for the University of Cincinnati. The system provided interactive map-based wayfinding across 46 campus buildings, turn-by-turn routing instructions, walk time estimates, and a dedicated ADA-accessible routing mode that restricted paths to elevator- and ramp-equipped corridors. The application was built using React, TypeScript, and MapLibre GL for the web platform, and was wrapped in Capacitor for native iOS and Android deployment, delivering a unified cross-platform experience with no dependency on proprietary mapping SDKs. Routing was powered by a custom Dijkstra implementation operating on a GeoJSON navigation graph of the UC campus. The project addressed a gap left by mainstream tools, namely the absence of ADA-aware campus routing and turn-by-turn indoor/outdoor wayfinding for University of Cincinnati students, faculty, and visitors.

NKU

Title: Knowledge Distillation from Large Reasoning Models to Compact Student Language Models

Team Members:  Gaurab Baral, Aaditya Khanal

Faculty Advisor: Dr. Yangyang Tao

Abstract: This project explores knowledge distillation from the large reasoning model DeepSeek-R1 to the compact Qwen2.5-7B using problems from the John O’Bryan Mathematics Competition (2011–2025). A Chain-of-Thought dataset was generated through a dual-agent framework and used to fine-tune the student model via LoRA on Apple Silicon with the MLX framework. Early stopping at 200 iterations mitigated overfitting, improving accuracy from 64.67% to 69.43%. The fine-tuned model also demonstrated strong generalization, achieving 73.1% accuracy on the MATH-500 benchmark. A key advantage is that the distilled Qwen2.5-7B model can be deployed on devices such as mobile phones or Raspberry Pi systems, enabling offline mathematical reasoning without internet access.