Towards online and personalized daily activity recognition, habit modeling, and anomaly detection for the solitary elderly through unobtrusive sensing

Lei Meng,Chunyan Miao,Cyril Leung

DOI: https://doi.org/10.1007/s11042-016-3267-8

IF: 2.577

2016-01-01

Multimedia Tools and Applications

Abstract:Rapid population aging and advances in sensing technologies motivate the development of unobtrusive healthcare systems, designed to unobtrusively collect an elderly’s personalized information of daily living and help him actively enjoy a healthy lifestyle. Existing studies towards this goal typically focus on recognition of activities of daily living (ADLs) and abnormal behavior detection. However, the applicability of these approaches is often limited by an offline analysis strategy, complex parameter tuning, obtrusive data collection, and a need for training data. To overcome these shortcomings, this paper presents a novel framework, named the online daily habit modeling and anomaly detection (ODHMAD) model, for the real-time personalized ADL recognition, habit modeling, and anomaly detection for the solitary elderly. In contrast to most existing studies which consider activity recognition and abnormal behavior detection separately, ODHMAD links both in a system. Specifically, ODHMAD performs online recognition of the elderly’s daily activities and dynamically models the elderly’s daily habit. In this way, ODHMAD recognizes the personalized abnormal behavior of an elderly by detecting anomalies in his learnt daily habit. The developed online activity recognition (OAR) algorithm determines the occurrence of activities by modeling the activation status of sensors. It has advantages of online learning, light parameter tuning, and no training data required. Moreover, OAR is able to obtain details of the detected activities. Experimental results demonstrate the effectiveness of the proposed OAR model for online activity recognition in terms of precision, false alarm rate, and miss detection rate.