University of North Dakota
A Privacy-Preserving, AI-Driven Framework For Bathroom Activity Monitoring And Health Recommendation In Elderly Care
Abstract
dc:description.abstractThe rapid aging of the global population necessitates scalable, non-invasive solutions, for monitoring Activities of Daily Living (ADLs) to detect early signs of health decline. This thesis presents a novel, end-to-end framework for monitoring bathroom activities—specifically bathing, toileting, and oral hygiene—using wearable sensors, machine learning and Agentic Artificial Intelligence (AI). The research addresses three critical gaps: the lack of granular activity recognition in privacy-sensitive spaces, the need for personalized health insights without manual logging, and the cybersecurity risks inherent in deploying AI in healthcare. First, algorithms utilizing wearable humidity and accelerometer sensor data were developed to classify hygiene behaviors with high precision. Second, a cross-platform mobile application was engineered to log these events and support a Retrieval-Augmented Generation (RAG) recommender system as proof of concept. Finally, a “dual-agent guardrail” security framework was implemented to protect against prompt injection and jailbreaks. This work demonstrates that secure, autonomous health monitoring is achievable using commodity hardware like the CarePredict Tempo device, offering viable pathway for aging-in-place technologies.
Degree
thesis:*- Name thesis:degree_name
- Master of Science (MS)
- Level thesis:degree_level
- Thesis
- Discipline thesis:degree_discipline
- Biomedical Engineering
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Obeng, Collins Poku
- Contributors dc:contributor
-
- Ryan Striker
- Enrique Alvarez Vazquez
Subjects
dc:subject × 1Identifiers
dc:identifier.*- Repository record dc:identifier
- https://commons.und.edu/theses/13967