Sensor-based activity recognition in the context of ambient assisted living systems: A review

Article type: Research Article

Authors: Patel, Ashish^{; *} | Shah, Jigarkumar

Affiliations: School of Technology, Pandit Deendayal Petroleum University, Gandhinagar, India. E-mails: ashish.patel@svmit.ac.in, jigarkumar.shah@sot.pdpu.ac.in

Correspondence: [*] Corresponding author. E-mail: ashish.patel@svmit.ac.in.

Abstract: Activity and behaviour monitoring of inhabitants play an essential role in an ambient environment. Different researchers proposed many promising solutions; this discipline, however, needs more accurate results. The main reason for this insufficiency is the imprecision data generated by the intelligent algorithms due to irregularities and unpredictability involved in users action and behaviour. The success of today’s system primarily ends on controlled and well-defined activities and conduct. However, in the real-world scenario, this is difficult to accomplish and requires more sophisticated strategies. Self-learning systems, in most cases, used to develop state-of-the-art in smart and cognitive environments. In this paper, we present a comprehensive review of different activity and behaviour analysis methods to identify seventeen critical challenges allied to ambient assisted living systems (AALs). Our primary objective is to offer a comprehensive guide to select the best approach to determine activity and human behaviour in the smart environment. Moreover, the study will present a better understanding of existing problems and a direction for future research.

Keywords: Activity recognition, ambient intelligence, assisted living, anomaly detection, smart environment

DOI: 10.3233/AIS-190529

Journal: Journal of Ambient Intelligence and Smart Environments, vol. 11, no. 4, pp. 301-322, 2019