Recent Advances in Human Activity Recognition for Healthcare (2019-2025)
Authors: Yu, H., He, Z., Wang, Q., Wang, Y., Cang, S., Linh, N.V.
Journal: Proceedings of Isoirs 2026 Moving Towards Embodied Intelligence in the AI Age 2026 6th International Symposium on Intelligent Robotics and Systems
Publication Date: 01/01/2026
DOI: 10.1109/ISOIRS70157.2026.11545278
Abstract:This paper reviews advances from 2019 to 2025, with emphasis on health-centric and real-world deployment. We synthesise progress across key sensing modalities-including wearable, ambient sensors, and hybrid systems-alongside developments in deep learning, federated learning, privacy-preserving analytics, and multimodal fusion. Also, the review examines clinical relevance and translational readiness. Major application domains include ageing-in-place, fall detection, rehabilitation, chronic disease management, clinical workflow analysis, and safety-critical monitoring. We further discuss evaluation practices, dataset limitations, generalisation challenges, and ethical considerations, particularly privacy, transparency, and fairness. By consolidating fragmented advances across wearable and ambient human activity recognition (HAR), this review identifies converging trends toward human-centred, unobtrusive, and context-aware healthcare systems, and outlines open challenges and future research directions. It serves as both an entry point for new researchers and a practical reference for developing deployable, trustworthy HAR solutions.
Source: Scopus