Sleep is essential for the maintenance of human life. Polysomnography (PSG) is a common method for the evaluation of sleep quality. However, the test requires attachment of many electrodes to the subject, and the constraint may cause great stress. Therefore, it is important to find a new method to free the subject from being burdened by electrodes;that is, to measure in an unconstraint and noncontact manner. In this study, we propose a new sleep measurement system using microwave sensing. The sensor detects body movements and respiratory movements in a noncontact manner. In this study, we analyzed sleep stage using PSG and simultaneously measured body movements and respiratory movements using a radio-frequency sensor. The mean and standard deviation of the number of body movements decreased, and the movements also became micro-movements when the degree of awakening decreased. The mean respiratory frequency did not change markedly in various sleep stages and showed little dispersion. The mean and standard deviation of the respiratory amplitude decreased as was observed for body movement, and the respiratory amplitude was stable when the degree of awakening decreased. We used the linear discriminant function to classify the sleep stages into:Wake, REM, Light and Deep sleep. The mean agreement rate was approximately 78% for four stages in total. Our results show that the sleep stages can be successfully estimated by body and respiratory movements during sleep.
