Research On Wearable Sleep Apnea Syndrome Detection System

Sleep Apnea Syndrome is a common sleep disease that directly causes problems such as hypertension,coronary heart disease,stroke and sudden death,which is a serious threat to people’s health.At present,the detection standard of sleep apnea syndrome is that patients are monitored by polysomnography in sleep laboratory,which is uncomfortable,complex and expensive,so it can not meet the needs of home sleep monitoring.Therefore,the development of a wearable,low cost,comfortable,and accurate Sleep Apnea Syndrome detection system has a certain research significance.This thesis combines ECG signals and respiratory signals is used to detect Sleep Apnea Syndrome and determine its symptoms.The wearable terminal part of the system uses flexible fabric electrodes to collect ECG signals,and through a breathing coil made of a polyvinylidene fluoride piezoelectric film to collect breathes signals,achieving wearability and improving comfort of the device.At the same time,the conditioning circuits for ECG signals and respiratory signals are designed respectively,after the signal is modulated,it is processed by the microcontroller and transmitted to the upper computer through Bluetooth.Two key algorithms have been developed on the upper computer.For ECG signals,an ECG-SAS detection algorithm based on support vector machines has been designed.First,this algorithm has de-noised ECG signals and extracted QRS waves,and then performed HRV time-frequency domain analysis,selected 12 time-frequency domain features for detecting SAS,then processed the features,and used the Support Vector Machine method to classify the features,and verified the classification model through the Apnea-ECG database.Finally,97.42% of the training set accuracy and 88.24% of the test set accuracy have been obtained.After the accurate detection of Sleep Apnea Syndrome has been achieved,this thesis has designed an algorithm for the incidence degree of Sleep Apnea Syndrome based on the respiratory signal.The algorithm first has preprocessed the respiratory signal by cubic spline interpolation,then extracted the breathing rate and obtains the number of sleep apnea.Then,the disease degree is simply judged according to the clinical diagnosis criteria.In this thesis,10 volunteers have been selected to verify the designed system,including the accuracy and sensitivity test of wearable terminal acquisition,the ECG signal preprocessing algorithm test of the upper computer,EEG-SAS detection algorithm test,and algorithm test of SAS incidence degree.The test results have showed that the wearable terminal can accurately and effectively collect ECG and respiratory signal,the recognition accuracy of EEG-SAS classification model is 83.72%,the error of sleep apnea times of the SAS incidence degree algorithm is between-2 and 3 times,which can meet the needs of the detection of sleep apnea syndrome and the extent of its onset.The detection system designed in this thesis has high accuracy,the system terminal is compact,light and comfortable to wear,which can realize long-term sleep monitoring.it is suitable for detection of sleep apnea syndrome,continuous collection of physical signs data,dynamic monitoring and assessment of sleep quality,disease warning and so on.