A Review on Sensor-based HAR Models Using GNN: AI in Healthcare
Nisha Dangol, Sazia Mahfuz · Procedia Computer Science · 2025
Graph Neural Networks (GNNs) have emerged as a transformative force in sensor-based Human Activity Recognition (HAR) by capturing complex spatial and temporal dependencies. In the era of Healthcare 5.0, where AI-driven workspaces and sustainable healthcare converge, the application of GNNs for HAR holds promise for enhancing patient monitoring, fall detection and personalized care. This paper systematically reviews recent advancements in GNN models specifically tailored for sensor-based HAR, examining methodologies from data acquisition to activity classification. By aligning these technological advances with AI-driven healthcare objectives, the review aims to outline the benefits, challenges and future directions of integrating GNNs into sustainable healthcare ecosystems.