Seite - 26 - in Intelligent Environments 2019 - Workshop Proceedings of the 15th International Conference on Intelligent Environments

Multisensor Monitoring System to Establish Correlations Between Sleep Performance and Environment Data Celestino GONÇALVESa,b, Fábio SILVAb,c, Paulo NOVAISb and Cesar ANALIDEb a Polytechnic Institute of Guarda, Guarda, Portugal c School of Management and Technology, Polytechnic Institute of Porto, Portugal b Algoritmi Centre, University of Minho, Braga, Portugal celestin@ipg.pt, {fabiosilva, pjon, analide}@di.uminho.pt Abstract. The importance of sleep in people's lives and concern about their quality have both been gaining an increasing relevance and awareness. Nowadays, the number of people who use mobile and wearable devices to monitor their sleep and day activity is quite revealing. But while some of those recent devices may even provide reliable measurements of some of the sleep parameters and sleep structure itself, they do not provide yet a justification for what has led to that situation and to those values. With the present study, we intend to verify and establish relationships between some environmental factors, such as temperature, humidity, luminosity, noise or air quality and between sleep performance and activity during the day. Several time series machine learning models were used to predict sleep stages and to estimate the level of day activity of a person. Keywords. Sleep monitoring, sleep structure, monitoring system, machine learning, sleep stages prediction, activity estimation 1. Introduction Sleep represents a significant part of human life, since about one-third is, or should be, spent sleeping, being an essential physiological function for the proper functioning of the brain [1], and therefore for the general well-being and for a good quality of life [2]. On the other hand, it is a complex process [3] that has only recently been given due attention by science and medicine, given the growing awareness of its importance [4]. It is a complex process and it can be measured considering multiple objective and subjective aspects and dimensions. For example, Buysse [4] proposes the following five dimensions to characterize sleep health: satisfaction, alertness, timing, efficiency, and duration. Examples of subjective methods for measuring sleep are the Pittsburgh Sleep Quality Index (PSQI) and sleep diaries, while the most commonly used objective methods are polysomnography (PSG) and actigraphy [2]. PSG is considered the ground- truth for measuring sleep, as it combines multiple sensors to measure various parameters, such as eye movement, brain activity, heart rate, muscle tone, and physical movement, however it is an expensive and intrusive method, usually carried out in clinical or laboratory environment, and is not normally used for data collection for very long periods. The actigraphy consists of the use of a device like a wrist watch, which by means of an accelerometer, measures movement as an indicator of wakefulness. It is a more Intelligent Environments 2019 A. Muñoz et al. (Eds.) © 2019 The authors and IOS Press. This article is published online with Open Access by IOS Press and distributed under the terms of the Creative Commons Attribution Non-Commercial License 4.0 (CC BY-NC 4.0). doi:10.3233/AISE190019 26