Research On Unobtrusive Sleep-disordered Breathing Detection And Analysis

Sleep is an important physiological phenomena,which is essential for human life.A good quality sleep is conducive to our health physically and mentally.Considering the development of human society and the changes in life style,our living conditions,which are full of higher stress and faster pace,result in plenty of sleep disorders,and among them,sleep apnea syndrome(SAS)is one of the most common ones.In addition,for the moment,existing medical care systems still can hardly provide continuous health monitoring and protection in home environment.As a result,making use of the widely-employed sensors to monitor night sleep and find out the healthy issues has realistic and instructive significance.From the thought of sleep monitoring in home perspective,we propose an early detection and severity analysis method for SAS based on ballistocardiogram(BCG).Patients themselves and their stakeholders can benefit from our work to recognize SAS at an early stage and know its severity,thus save valuable time for medical treatment.This work could be supposed as a significative trial for health assistant in home environment.This article mainly focuses on the following four detailed studies:(1)A non-intrusive BCG signal acquisition method is discussed and presented based on the micro-movement sensitive mattress(MSM).The method can sense the signal without any electrodes attached on human body,has no interference on human normal sleep process,and is suitable for generalizing in home.(2)We propose a BCG signal preprocessing method in this work by using wavelet decomposition.With the fixed-width sliding window,the heartbeat interval time sequences are calculated,moreover,in regarding with the existing missed and spurious RR intervals,we recommend a versatile correction algorithm.(3)Take full advantage of the BCG inherent and rich content,i.e.the heartbeat interval information and breathing-related information.Meanwhile,the experiment dataset attributes and demographic indices are also extracted for fine-grained SAS early detection,in which the sleep stages are viewed as the basis.(4)We introduce an automated severity measurement analysis for SAS from the sleep-related breathing events(SBEs)point of view,the proposed method consists of three main sub-procedures,namely the SBEs detection,the SBEs identification and the severity evaluation.Our method can detect and screen the SBEs effectively,and also can get comparatively accurate severity measurement results.It is important to mention that we collect some real-world clinical sleep data as our experiment dataset,and the experimental results validate the effectiveness and reliability of our work in this research.