Early online detection of upper airway obstructions in obstructive sleep apnoea syndrome (OSAS) patients

The obstructive sleep apnoea syndrome (OSAS) is a diagnosis related to snoring and caused by a collapse in the upper airway. OSAS patients suffer from desaturated oxygen levels during sleep as well as daytime sleepiness. In this paper, we propose a system able to identify and detect respiratory disorders online based on monitoring the airflow amplitude from a sleeping OSAS patient. By the use of chi(2)-analysis and a Haar wavelet transform on signals performed offline, reference templates indicating the specific apnoea pattern for four different patients are constructed and used for similarity matching against online signals. Detection is performed in the early stages of an upcoming airway dysfunction, thus providing an opportunity to alert the patient at sleep. The system-testing results indicate robust performance and flexibility for the patient. Our proposed solution can in turn operate as an alternative to today's OSAS treatment of choice, the continuous positive airway pressure (CPAP).