Available online at www.sciencedirect.com Procedia Engineering 41 (2012) 1128–1134 International Symposium on Robotics and Intelligent Sensors 2012 (IRIS 2012) An Intelligent Sensing System for Sleep Motion and Stage Analysis Jin U Bak, Nikolaos Giakoumidis, Gagyung Kim, Haiwei Dong, Nikolaos Mavridis* New York University Abu Dhabi, P.O. Box 129188, Abu Dhabi, United Arab Emirates Abstract Monitoring the movements of the human during sleep can potentially give us a good estimation of aspects of the bodily as well as mental state of a human. When such data are combined, either with the knowledge of a sleep pathologist or with a special automated diagnosis system, they could prove quite useful towards the diagnosis of various types of sleep disorders such as parasomnias, insomnia, and dyspnea. Furthermore, such data could also be useful towards diagnosis of various medical conditions, and towards quantitative evaluation of the effects of drug therapy that is administered to a patient who is suffering from poor sleep quality, an important indication of which is the duration and patterns of various sleep stages. The intelligent sensing system that we present consists of a thermal infrared camera, a budget three-electrode budget EEG device, and algorithms for analysis and motion processing which we designed for this system. The main measurables that we derived from our system are of three kinds: a) descriptions of sleep stages (personalized probabilistic model), b) movement graphs, and c) relations between stages and motion. An empirical study with two subjects was carried out, where sensory recordings for multiple nights were captured and analyzed, illustrating the capacity of our sensory system towards providing the above measurables, and quite importantly, towards acting as a strong foundation for future wider deployment of in-home sleep self-monitoring and diagnosis tools. © 2012 The Authors. Published by Elsevier Ltd. Selection and/or peer-review under responsibility of the Centre of Humanoid Robots and Bio-Sensor (HuRoBs), Faculty of Mechanical Engineering, Universiti Teknologi MARA. Keywords: Intelligent sensors; sleep disorders; sleep stages; sleep movement; computer vision 1. Introduction Assuming that people sleep 8 hours a day on average, people spend about 33% of their lives sleeping. Most importantly, getting a good sleep every night is highly significant as it plays a key role in providing quality time while a person is awake. Studies on sleep have been done in many different ways such as EEG monitoring, body movement tracking using pressure mat [1] or bed temperature monitoring [2]. Sleep monitoring could be crucial in detecting sleep disorders and treatment of sleep disorders [3]. In order to assess the quality of sleep, identifying and measuring the sleep stages that a person is quite useful, as there is a strong correlation between sleep quality and sleep stage duration and patterns. Sleep stages can be subdivided in Non-Rapid Eye Movement (NREM) and Rapid Eye Movement (REM) sleep. People go through what is called as ‘Sleep Cycle’ that consists of combination of periodic NREM and REM sleep. NREM sleep can be further divided into four stages; N1, N2, N3 and N4 [5]. Sleep stages are traditionally identified using EEG by collecting surface electrical signals arising from internal brain activity, and recognizing the appropriate sleep stage depending on the type and bare frequency of the wave form that arises. Furthermore, monitoring sleep through traditional multi-electrode hospital EEG devices involves direct contact between the user’s head and multiple electrodes, which could create discomfort. On the other hand, apart from * Corresponding author. E-mail address: nikolaos.mavridis@nyu.edu.