Non-invasive wearable sensing system for sleep disorder monitoring

• Autor(a) principal: Alberto, Mauro Cruz
• Data de Publicação: 2017
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Texto Completo: http://hdl.handle.net/10400.1/10121
• topic: Distúrbios de sono; Apneia; Monitoramento remoto da saúde

This Master Thesis introduced a proposal of a remote sensory system for the detection of sleep disorders in geriatric outpatients. Although the most accurate solution would be an in-depth study in a sleep clinic, it is not a realistic environment for the elderly. The objective is that the patient stays at home, and without changing their daily routines, the clinicians get objective information in order to make a correct diagnosis of the sleep disorders. Sleep disorders are often classified as medical disorders corresponding to modifications on the sleep patterns and the amount of these modifications increase with age. However, regularly, these illnesses are undiagnosed, since is hard for the patients to explain the symptoms to the doctor. To achieve the proposed objective, we studied the polysomnography bio-signals that could be used to accurate reflect the sleep disorders occurrences. We designed a Body Sensor Network (BSN) to be divided into both movement assessment (Accelerometer and Gyroscope) and biomedical signals (EMG, ECG, PPG, GSR) evaluation. These signals, reflecting both breathing and cardiac activities, are processed by a specifically developed algorithm. The reduction of the number of sensors was also envisaged, and it was decided to use 3 biomedical sensors instead of the minimum of 22 sensors used by polysomnography. Thus, to offer better visualization of the recorded signals a software interface was developed to include the processing and visualization of the signals. To identify the sleep stage and apnea state, we settled an algorithm that processes both ECG and EMG. To validate this algorithm, it was decided to use two sources of data: PhysioNet data base containing ECG and EMG signals and data recorded by our BSN on volunteers. With this work, we were able to build a BSN capable of detecting a set of sleep disorders, without using any invasive method. The network provides reliable data, and using the developed interface, it helps elderly health providers to carry out an in-depth analysis of the information and to better identify sleep disorders.