「世大智科/天才家居」-我們創業囉
Contact Professor: Yeh-Liang Hsu (徐業良)

活動感知床墊機器學習睡眠演算法開發及物聯網應用

睡眠資訊是判斷人體生理健康狀況的重要資訊之一。睡眠多項生理檢查儀器(polysomnography, PSG)是醫學界所依賴的睡眠監測標準設備,然而使用PSG時需在特定的睡眠實驗中心進行,較不適合在居家環境下進行長期監測。因此許多研究開始發展可攜式穿戴裝置、非察覺式的臥床監測技術,讓受試者不必依賴專業睡眠監測設備也能量取生理訊號、偵測睡眠品質。 本研究主要目的是活動感知床墊WhizPad的商品化與睡眠演算法之開發,在商品化過程中,更換適合物聯網的晶片於WhizPad硬體外,也在軟體方面重新測試WhizPad的基礎感知功能演算法,包含臥離床判別及活動偵測判別,完成一套可偵測人體活動的睡眠監測系統,記錄使用者的睡眠歷程。最後,本研究以此系統所使用WhizPad與PSG同時進行睡眠實驗,比對WhizPad與PSG的睡眠狀態,應用了機器學習方法建立睡眠判讀模型並編譯於控制晶片中,期望提供一個能在居家環境下睡眠品質偵測的商品化物聯網床墊。

Development of the Machine Learning Algorism of Sleep Status for Motion Sensing Mattress and IoT Applications

Sleep is one of the important information to measure the health of human physiology. Polysomnography(PSG)is the current golden standard equipment for measuring sleep. However, this professional sleep monitoring equipment should be used only on specific laboratory—a controlled base under a sleep technician. It’s hard to maintain and have long-term detecting in home environment. So, many academic start to develop portable wearing device、unobtrusive sensing technology in bed. The subjects do not have to rely on the professional sleep monitoring equipment can still measuring physiologic signals and detecting sleep quality. The purpose of this research is using commercialized motion sensing mattress to develop the algorism of sleep status versus PSG. In the process of commercialization, we not only replaced the original microprocessor with IoT microprocessor fot the WhizPad hardware device, but build several basic features algorithm on software, including on/off bed feature discriminant and activity awareness discriminant. This is an unobtrusive sleep monitoring system which can detecting physiologic parameters during sleep then recording. At last, WhizPad and polysomnography were simultaneously collected data during sleep experiment, then compared with each other. Using machine learning to build sleep status validation module and program it into microprocessor. We hope provide a special commercialized IoT mattress which can detecting physiologic parameters、measuring sleep quality in home environment.
Academic Thesis
Last Updated:2017/8/1