Non-EEG Sleep Monitor System And Algorithm Research

Sleep disorders seriously affect people’s lives. It is important for preventing and diagnosising sleep related diseases to do sleep detection early. The standard of sleep staging criteria is Polysomnography which depends on EEG is, the standard of sleep staging criteria. While there are some problems using PSG for sleep monitoring,such as It is compelted to operate and may cause feel uncomfortable. Therefore, non-EEG sleep monitoring will be the development tendency of future daily sleep monitoring using the ECG, EOG, EMG and other signals. The present non-EEG sleep monitoring systems have serveral problems, like unspecific sampling information, low sleep staging accuracy and can’t be used to clinical diagnosis.Paper developed a sleep monitor system which realize two kinds of sleep monitor modes: EEG included sleep monitor named PSG detecting and non-EEG sleep monitor using EEG, EMG, breathing, pulse.Paper research content is as listed:1, Based on cortex-M3 and highly, integrated hardware platform, designed a multi-function, high performance embedded software. System relized two sleep monitor modes, and was able to realize multi-channel physiological signals sampling with high accuracy undering low CPU frequency, and was able to store and transmit a large amount of data and so on.2, Paper proposed a kind of non-EEG sleep staging algorithm with EOG and EMG. Using zero phase digital filter for signals processing and extracting features in time domain. The doing adaptive threshold extracting and edges detecting for features, and classifying signals with multi-features fusion method. Finally, Based on PSG and combined with signal classification results, studied the sleep staging method with EOG and EMG, and finished the non-EEG sleep stage classification.3, Testing the algorithm using 50 pieces of data from the international sleep database. The results show that the accuracy in detecting the length of Wake, light sleep, deep sleep and REM is higher than 73%. At the same time, Applied algorithm to the non-EEG sleep monitoring system, and do experiments in clinical.The algorithm is based on physiological characteristics, and the signals chosen has specific physical characteristics. It has the advantage of high accuracy superior to the existing non-EEG sleep staging methods, and has important practical application values on the fields of household medical, clinical diagnosis and special field sleep monitoring.