AUTOMATED OBSTRUCTIVE SLEEP APNEA (OSA) EVENTS CLASSIFICATION BY EFFECTIVE RADAR CROSS SECTION METHOD

Abstract

A sleep disorder is a medical disorder of the sleep patterns of a person which involves problem with quality, timing and amount of sleep. Obstructive Sleep Apnea (OSA) is one of the most common sleep disorders. OSA is briefly and repeatedly cessation of breathing when throat muscles intermittently relax and block the airway during sleep. Almost 22 million Americans suffer from sleep apnea, with 80 percent of the cases of moderate and severe obstructive sleep apnea undiagnosed. Often sleep disorder go undiagnosed as its diagnosis process is difficult and costly. Polysomnography (PSG) is considered as gold standard test for detecting sleep disorder. During PSG, different sensors are attached to suspected patient and performed in specific sleep clinic overnight in presence of a sleep technician. The patient must sleep in the clinic with different sensors attached to his body which is very uncomfortable. Multi-night PSG tests are rarely performed despite large night to night variance in sleep outcomes for those with OSA. Consequently, numerous alternative sleep monitoring technologies have been developed to overcome the disadvantages of full night PSG recording, using reduced number of sensors and allowing for at-home recording for natural sleep conditions of patient. In-home sleep monitoring system using Microwave Doppler radar is gaining attention as it is unobtrusive and non-contact form of measurement. Most of the reported results in literature focused on utilizing radar-reflected signal amplitude to recognize Obstructive sleep apnea (OSA) events which requires iterative analysis and cannot recommend about sleep positions also (supine, prone and side). In this dissertation, we propose a new, robust and automated ERCS-based (Effective Radar Cross section) method for classifying OSA events (normal, apnea and hypopnea) by integrating radar system in a clinical setup. In prior attempt, ERCS has been proven versatile method to recognize different sleep postures. Here, two different machine learning classifiers (K-nearest neighbor (KNN) and Support Vector machine (SVM) is employed to recognize OSA events from radar captured ERCS and breathing rate measurement from five different patients’ clinical study. The proposed system has several potential applications in healthcare, continuous monitoring and security/surveillance applications.