睡眠在人的生活中扮演非常重要的角色,對於生理及心理的狀態恢復有一定的作用,因此自身的睡眠品質已被漸漸的重視。然而一般的穿戴式睡眠裝置礙於硬體限制,演算法設計無法太過複雜,進而影響評估睡眠品質的準確性。因此本論文提出一套新的系統,主要應用於穿戴式裝置,其中包含了收集心率參數的前端裝置,在受測者睡眠期間紀錄資料。此外論文與安卓系統結合,以簡單的介面幫助受測者自身的睡眠階段變化情況。考慮到穿戴式裝置硬體的限制,演算法設計參考心跳變異率的時域分析,參考SDNN(Standard Deviation of all Normal to Normal Intervals)時域分析作為主要判斷準則,最後再根據三軸加速度值的變化,綜合整理出完整的睡眠品質分析。 最後系統與MIT-BIH(Polysomnographic Database)睡眠資料庫驗證,整體睡眠階段命中率達70%以上;評估結果與其他系統比較下,MAPE(Mean Absolute Percentage Error)的結果分別為0.082%及0.1365%。相較其他穿戴式睡眠裝置,能提供更準確的睡眠資訊。
Sleeping is important in people’s life, which is occupy one-third of life. People have to go to the hospital to detect the sleep disorder when something bad during sleep. Due to the complex procedure, we expect to develop a new system to detect sleep quality, which used few physiological parameters. Considering the limitations of wearable devices, the algorithms using relatively simple time-domain analysis. In addition, the system will combine with android environment to help people to understand self-condition easily. The overall hit rate was above 70% of sleep stage detection which was compared to MIT-BIH sleep database. And the MAPE of sleep efficiency verification with other system were 0.082% and 0.1365%. Compared to other systems, it can offer more accurate information.