Edge AI and Wearable Sensor Integration for Fall Detection in Elderly Independent Living

Rajan, Santhosh (2025) Edge AI and Wearable Sensor Integration for Fall Detection in Elderly Independent Living. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (837kB) | Preview

Abstract

The trend of an ageing population is intensifying the need for affordable, privacy-oriented, and ample care systems that support longevity and independence for our elderly adult population. The research in this paper presents a modular framework for an AI-enabled virtual assistant to detect falls and observe the physical activities of seniors who live alone. The system processes multimodal data from wearable accelerometers and gyroscopes, using benchmark datasets like MHealth, WISDM, and SisFall to train and evaluate supervised machine learning models, specifically Random Forests and Long Short Term Memory (LSTM) networks, that were assessed for their potential to monitor falls in real time, with preliminary results indicating modest ( 31%) accuracy in this iteration.

This project conceptualized several implementation methods and used both centralized and edge computing configurations, but no practical experimentation was done, only in simulation. The centralized approach is proposed to exploit a dedicated server to optimize higher valued computations, especially with higher amounts of data transfers and processing. The edge system, proposed as an alternative, would limit data transfer requirements, and build models locally in each client node using federated learning, which is suggested as a privacy-preserving strategy but was not implemented in this study. The overall design also included the possibility of a voice interface, posed as an option with automatic speech recognition to provide better access for those with mobility issues. The proposed design represents a conceptual compromise between accuracy and responsiveness of detection, as well as user independence, and may serve as a potential pathway in exploring AI approaches to elderly care solutions in real-world settings.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Thomas, Lavish UNSPECIFIED
Uncontrolled Keywords:	Elderly care; AI-powered assistants; fall detection; wearable sensors; IoT health monitoring; activity recognition; edge computing; federated learning; privacy-preserving AI; assistive technology
Subjects:	Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence Q Science > QA Mathematics > Electronic computers. Computer science > Computer Systems > Computers > Electronic data processing > Electronic data processing--Distributed processing > Edge computing T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science > Computer Systems > Computers > Electronic data processing > Electronic data processing--Distributed processing > Edge computing R Medicine > Healthcare Industry T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Telecommunications > Computer networks > Internet of things
Divisions:	School of Computing > Master of Science in Artificial Intelligence
Depositing User:	Ciara O'Brien
Date Deposited:	04 Jun 2026 15:01
Last Modified:	04 Jun 2026 15:01
URI:	https://norma.ncirl.ie/id/eprint/9340

Actions (login required)

View Item