What follows is the lab's working portfolio. Active projects are the work in motion, often with outside collaborators. Completed projects are software we have released and the published or presented work that came out of independent studies and undergraduate research.
What we are building right now, with collaborators and students.
An AI-powered math tutoring assistant built on the historic Mathforum problem-thread archive. The project explores RAG (retrieval-augmented generation) and fine-tuning strategies to provide pedagogically sound feedback on student mathematical thinking, in collaboration with Jason Silverman and the Gates Foundation.
A full-stack wearable system for continuous, ambulatory cardiorespiratory monitoring. The project integrates signal acquisition, edge processing, and machine learning inference for real-time health monitoring outside clinical settings.
A reconfigurable wireless testbed at the Drexel Wireless Systems Lab supporting experiments in RF sensing, smart grid communications, and spectrum management.
End-to-end paper on passive wearable respiratory device design and the analysis algorithms that monitor respiratory activity, covering hardware, signal processing, and clinical validation.
Research into intelligent antenna state selection for RFID-based sensing, targeting improved read range and accuracy in multi-tag environments.
Software we have released, papers we have written, and the undergraduate work behind much of it.
An interactive, no-code neural network visualizer that lets users upload data, define a network architecture, train a model, and explore its behavior through Sankey diagram visualizations — all in the browser. Developed by Ursinus students Kevin Hoffman and Michael Cummins. Source on GitHub.
A serverless, browser-based development environment for student practice and rapid instructor exercise development, supporting multiple languages with no setup required. Developed with Christopher Tralie.
A platform to collect data from IoT-based sensors — including knitted smart textiles — for HIPAA-compliant processing and research. Developed with the Drexel Wireless Systems Lab.
Integrates with the IoT Sensor Framework to provide real-time or offline signal processing and machine learning on data collected by the sensor framework. Developed with the Drexel Wireless Systems Lab.
Student research comparing retrieval-augmented generation and fine-tuning for building an AI math tutoring assistant on the Mathforum archive. Presented at Ursinus CoSA 2026. Developed by Michael Cummins, Matt Latta, and Claire Shoemaker.
A machine learning pipeline for minute-level and subject-level sleep apnea detection using ECG signals. Presented at Ursinus CoSA 2026. Developed by Andrei Bogdan.
Tools for making machine learning model decisions more interpretable and transparent. Presented at Ursinus CoSA 2025. Developed by Michael Cummins.
A suite of financial literacy and planning tools developed for the Grizzly Guide to Wealth platform. Presented at Ursinus CoSA 2025. Developed by Eugene Thompson and George Psaradakis.
A mobile campus orientation app using ArcGIS to help new students navigate Ursinus College. Presented at Ursinus CoSA 2023. Developed by Arthur Artene and Tristan Ashcroft.
A radio-frequency security layer for IoT sensor networks, adding authentication and encryption at the physical layer. Presented at Ursinus CoSA 2023. Developed by Jonas Ling.
Exploring algorithmic and neural network approaches to automatically detect and catalog craters on the lunar surface from LRO imagery. Presented at Ursinus CoSA 2023. Developed by Dylan Melby.
An AI-driven interactive theater performance exploring human-machine collaboration in live performance art. Presented at Ursinus CoSA 2026. Developed by Levi Fritz and Shannon Zura.
Honors and research independent studies investigating AI applications in educational contexts, including generative AI, classroom integration, and pedagogy. Conducted by four students across CS394 and the Honors thesis sequence (CS491–CS492).
Honors independent study designing an AI-assisted tool for automated Wi-Fi security assessment and penetration testing. Conducted in the Honors thesis sequence (CS491–CS492).
Research independent studies using RF signal analysis and machine learning to estimate human body pose without cameras. Conducted by two students (CS394).
Research independent studies in RFID-based indoor localization, multi-tag tracking, and dynamic radar sensing techniques. Conducted by four students across CS394.
Multi-semester research in RF-based biomedical signal acquisition and analysis using wearable IoT sensor devices. Conducted by four students across CS391 and CS394.
Research and educational module development for software-defined radio (SDR) platforms in smart-grid communication scenarios. Conducted by multiple students across CS392 and CS394.
Research independent studies developing methods to visualize internal feature representations in trained neural networks, connecting to broader explainable AI themes. Conducted by multiple students across CS391 and CS392.
Research independent study in natural language processing techniques and applications, including retrieval-augmented approaches. Conducted by three students (CS394).
Independent study developing an open financial literacy toolkit for educational and personal finance use. Conducted in CS392.
Research independent study in multi-sensor data fusion techniques for wearable and IoT applications. Conducted by two students (CS392).
Research independent study exploring side-channel electromagnetic signatures for hardware-level malware detection. Conducted in CS391.
Research independent study in virtual reality museum environment design and interactive digital exhibit development. Conducted in CS391.
Independent studies in reliable RF-based emergency communication systems and protocols, including amateur radio integration. Conducted across CS392 and CS394.
Research independent study applying deep learning techniques to astrophysical image and data analysis. Conducted in CS391.
