Analysis Of Cough And SaO2 Signals Based On The MSMSMS

The Micro Movement Sensitive Mattress Sleep Monitoring System (MSMSMS) can continuously measure respiratory wave, ballistocardiogram, body movement, position and posture for long time without sticking any electrode on patients' bodies. The thesis analyzed vibration and SaO2 signals of MSMSMS.It is very important to identify cough and snore when extracting respiratory events and analyzing sleep stage in sleep monitoring. Cough monitor can also help doctors to diagnose diseases. Most traditional cough monitor record audio signal which has large amount of data and the algorithm is too complex. High-pressure air in lung dashes out of glottis when coughing and causes unique vibration of chest wall. We added a special bone conduction whose center frequency is 20Hz and a voice sensor whose frequency ranges from 20Hz to 2 KHz to the MSMSMS to collect cough vibration and sound. The two signals are both transformed to 10bit digital signals at a sampling rate of 100Hz. As a contrast, audio record using an Mp3 is synchronously done. Low sampling rate makes it possible to be used to simultaneously monitor many patients in hospital. We studied slope, zero-across rate, declining time of the two signals and developed an algorithm to differentiate cough and snore. It could be used to monitor cough at nighttime and validate the results of respiratory events extraction.REM sleep plays an important role in sleep architecture. The thesis made a deep research on recognizing REM sleep based on SaO2 (oxygen saturation) recorded by the MSMSMS. After observing plenty of data, we find that REM sleep and declining of SaO2 are highly related, especially in patients suffering from middling and serious SAHS. The thesis developed an algorithm to automatically identify REM sleep based on vales of SaO2. First we identified SaO2 vales based on its mean and variation, validation methods were added after applied the algorithm to plenty of data, and then we found REM sleep recording to these vales and sleep analysis knowledge base. After validation by applying it to 30 patients' data and expert on sleep analysis from Chaoyang Hospital, it is proved that the algorithm meets the demand of clinic.The thesis analyzed vibration and SaO2 signals of MSMSMS and the MSMSMS could analyze sleep stage and diagnosis SAHS more accurately. This cough extracting algorithm could also be used in sleep holter and cough holter. The MSMSMS is a novel low-stress sleep monitor, physiological parameters recorded by which contain ample information.