| Summary: | 碩士 === 國立臺北科技大學 === 電機工程研究所 === 103 === Sleep is one of the most basic physiological need. The great quality of sleep is able to eliminate fatigue. Therefore, sleep quality estimation becomes very important. In the past, people mostly used polysomnography (PSG) to evaluate sleep stage. However, PSG had a number of disadvantages including too much wire, expensive cost, and complexity of computation, so that PSG isn’t suitable to home monitoring. As a result, there are many scholars dedicated themselves to researching sleep stage analysis using non-invasive sensor. However, most of these researches used threshold to analyze stage. Wherein the statuses of the conversion between each sleep stage are overlooked. For sleep stage estimation, such approach has its drawback.
Therefore, this paper proposed sleep stage analysis based on fuzzy inference and finite state machine, using Bluetooth heartrate monitor and force sensor to collect heartrate, heartrate variability and body movement during sleep. Further, this paper used fuzzy inference system to estimate sleep depth according to classification and characteristics of sleep proposed by American Academy of Sleep Medicine. Next, this paper used the sleep depth to determine sleep stage by finite state machine. To sum up, this study can analyze sleep quality according to the time distribution of sleep stage.
The method this paper proposed can effectively improve the irrational state change and the inflexible threshold. The system transmitted data to the server by means of Bluetooth heartrate monitor and force sensor via wireless transmission. This study not only reduced the burden of measurement, but also produced the similar result with PSG. Therefore, this system is easy to estimate the sleep stage at home. It will be less expensive and comfortable of use.
|
|---|
