Improving the validity of lifelogging physical activity measures in an internet of things environment

Abstract

Recently, the popular use of wearable devices and mobile apps makes the effectively capture of lifelogging physical activity data in an Internet of Things (IoT) environment possible. The effective collection of measures of physical activity in the long term is beneficial to interdisciplinary healthcare research and collaboration from clinicians, researchers to patients. However, due to heterogeneity of connected devices and rapid change of diverse life patterns in an IoT environment, lifelogging physical activity information captured by mobile devices usually contains much uncertainty. In this paper, we provide a comprehensive review of existing life-logging physical activity measurement devices, and identify regular and irregular uncertainties of these activity measures in an IoT environment. We then project the distribution of irregular uncertainty by defining a walking speed related score named as Daily Activity in Physical Space (DAPS). Finally, we present an ellipse fitting model based validity improvement method for reducing uncertainties of life-logging physical activity measures in an IoT environment. The experimental results reflect that the proposed method effectively improves the validity of physical activity measures in a healthcare platform.