Sleep Apnea Syndrome Screening System Based On Android Mobile Phone

With the increasing of the life pressure in modern society, more and more people suffering from sleep disorders, in which the most serious one is the sleep apnea syndrome. This disease generally occurs at night and is hard to be found. It caused great harm to human health, so the prevention and diagnosis of this disease is very important. Since the gold standard for the diagnosis of sleep apnea syndrome -- multichannel sleep instrument, can not be spread to families, currently this thesis designed a portable sleep apnea screening system, which is suitable for the family, and is able to do a preliminary screening of the disease without affecting the user's normal sleep. The main contents of this thesis are as follows:First of all, the sleep apnea syndrome screening system present by this thesis is to combine the breath signal, blood oxygen signal and snoring signal in order to make full use of their advantages for disease resolution and improves the screening capability of the whole system. The work of this system contains hardware design and software design. In the aspect of hardware, the breathing and oxygen signals are collected from thoracic and abdominal breathing zone and clip type oxygen probe, snoring is collected by the phone's own recording function. The acquisition equipment has reached a high degree of portability and comfort.Secondly, in the treatment of snoring signal endpoint detection, this thesis proposed a new endpoint detection algorithm called EEMD-Co. This method is suitable for the detection of snoring, which is a kind of nonlinear and non-stationary speech signal. The experimental results show that the accuracy of this method is higher than the commonly used endpoint detection algorithm under the same conditions.Finally, in aspect of the assistant diagnosis of sleep apnea syndrome, a new feature called pitch energy ratio is proposed in this thesis, which can effectively distinguish between patients with sleep apnea syndrome and simple snorer. The classification results through SVM training shows that the system has a high accuracy for the diagnosis of sleep apnea syndrome.