Title: Daily activity monitoring system designed for elderly people using hidden Markov models based on real world datasets

Authors: Chaima Bouali; Olivier Habert; Abderrahim Tahiri

Addresses: Laboratory of Information Systems and Software Engineering, National School of Applied Sciences, Tetouan, Morocco ' Laboratoire de Conception, Optimisation et Modélisation des Systèmes, Metz, France ' Laboratory of Information Systems and Software Engineering, National School of Applied Sciences, Tetouan, Morocco

Abstract: This work describes how our abnormal behaviour detection system functions for seniors in their home. Our research is based on the data gathered by a domotic box that is available for purchase. The box was initially intended to continuously detect the owners' daily actions using non-intrusive home automation sensors. The enhancement of the detection of the health changes, deducted by the abnormal behaviour of the user, is closely related to the evolution of the activity recognition of the box. Our system aims to report a relevant context-aware alert to health care service experts. By refining the detection of the activity level of the occupants, we could identify warning manifestations for early intervention. In this paper, we will describe the process of pointing out irregularity in the daily activity pattern of a user or the detection of a malfunction of the box to maintain the accuracy of the service it offers.

Keywords: activity monitoring; ambient assisted living; AAL; data processing; smart homes; elderly; assisted living; sensors; statistical approach.

DOI: 10.1504/IJISTA.2023.130555

International Journal of Intelligent Systems Technologies and Applications, 2023 Vol.21 No.1, pp.40 - 55

Received: 15 Dec 2021
Received in revised form: 19 Sep 2022
Accepted: 17 Oct 2022
Published online: 27 Apr 2023 *