請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/72875
標題: | 基於物聯網架構之環境光雲端運算平台 Cloud Computing Platform of Ambient Light Based on IoT Architecture |
作者: | Yi-Hsuan Chuang 莊易軒 |
指導教授: | 陳中平(Chung-Ping Chen) |
關鍵字: | 穿戴式,多通道,環境光,場域辨識,心理壓力, Wearable,Multi-channel,Ambient light,Field identification,Psychological stress, |
出版年 : | 2019 |
學位: | 碩士 |
摘要: | 隨著科技的發展,人造光的普及為生活上帶來了許多便利之處,但同時也直接或間接地導致人體晝夜節律的失調進而產生睡眠、心理壓力與情緒上的問題。近年來有越來越多研究探討環境光及人造光對於人的生理及心理產生的影響,然而卻缺乏能夠長時間持續記錄環境光成分的裝置,使得實驗設計大大的受限。
本研究除了整合實驗室先前關於腦波的研究之外,也開發了新型的穿戴式多通道環境光感測裝置與系統,實際測量之後,裝置使用3.7V/500mAh的鋰電池可以待機將近37小時,重量僅有28克比現有的光感測儀器輕9倍,體積僅為32.5立方公分,比現有的光感測儀器小6倍,只有相當於一個火柴盒大小,並且能夠收取從紫光(410 nm)到紅外光(940 nm)總共18種不同波長的環境光資訊。因此,相比於其他裝置,非常適合應用於需要長時間配戴的環境光採集任務當中。 此外,本研究將穿戴式多通道環境光感測裝置所採集的環境光資訊用於場域辨識任務中,驗證了裝置的可用性以及環境光資訊在機器學習上的可行性,最後使用SVM於場域辨識任務中達到95%的驗證與測試準確率。 本研究的最後還設計了一個實驗來分析心理壓力與環境光之間的關係,在10個案例中發現,在相同的生活壓力等級下,缺乏綠光(510nm、535nm)、紅光(680nm、705nm、730nm)和紅外光(760nm、810nm、900nm)的現象在自我評估為高度心理壓力的組別中很常見。 With the development of science and technology, the popularity of artificial light has brought many conveniences to life. However, it has also directly or indirectly led to the imbalance of the human circadian rhythm. In recent years, a growing number of studies have examined the effects of ambient light and artificial light on human physiology and psychology. However, there is a lack of devices capable of continuously recording ambient light components for a long period of time, which greatly limits the experimental design. In addition to integrating previous research on EEG, we also develop a new wearable multi-channel ambient light sensing device and system. After the actual measurement, the wearable multi-channel ambient light sensing device can stand by for nearly 37 hours with a 3.7V/500mAh lithium battery and weighs only 28g, which is 9-times lighter than traditional light sensing instruments. The volume is only 32.5cm3, which is 6-times smaller than traditional light sensing instruments and equivalent to a matchbox. The device can also collect a total of 18 kinds of ambient light information from violet (410nm) to infrared (940nm). Therefore, it is very suitable for ambient light data collection tasks that require long-term wearing compared to other devices. The environmental field identification is performed using the data collected by the wearable ambient light sensing device to verify the usability of the wearable ambient light sensing device and the availability of the ambient light in machine learning. Finally, we adopt SVM and 95% validation and test accuracy are achieved in the field identification task. In addition, an experiment is designed to analyze the relationship between psychological stress and ambient light. In total 10 cases here received, the high-stress group had lower exposure to green light (510nm, 535nm), red light (680nm, 705nm, 730nm), and infrared (760nm, 810nm, 900nm). |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/72875 |
DOI: | 10.6342/NTU201901435 |
全文授權: | 有償授權 |
顯示於系所單位: | 生醫電子與資訊學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-108-1.pdf 目前未授權公開取用 | 5.02 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。