睡眠是很重要的,人們經由睡眠得到適當的休息。但現代人因壓力或身體健康狀況不良等因素,造成睡眠品質不佳,發生睡眠障礙的情形越來越多。過去為了評估睡眠障礙的程度,主要是透過多項生理訊號儀器來輔助評估睡眠障礙,但是其價格昂貴且需要專業人員操作等缺點,並不適合應用於居家睡眠品質照護。 因此本文提出居家型品質監測系統,透過單一生理訊評估睡眠品質。系統是以非侵入式的方式量測心率訊號,並擷取心率頻譜特徵值,再透過具有學習能力且運算容易的灰色適應共振網路作睡眠等級分級,由睡眠的深淺及長短來分析睡眠品質。評估結果與標準睡眠資料庫相比,透過本系統所評估出來的正確率約可達到標準的七成。經由本系統,病人可以在家中進行睡眠品質的監測,透過初期篩選輔助評估病人是否患有睡眠障礙等問題,再逕行自醫院接受進一步治療。
Sleeping is very important to human being, because people keep life depend on sufficient sleeping time. Recently people have more pressure from living or worse healthy. In the past people used polysomnogram to estimate sleep disorder. But it has some drawbacks such as expensive and also hard to use. It is difficult to estimate the quality of sleeping in home care. In this thesis, we design and implement a sleeping quality monitor system for home care applications. It use a non-invasive heart rate measurement and a classificatory algorithm GreyART. The advantages of GreyART are low computation and ability of learning. The system classifies the stage of sleeping and also analysis the quality of sleeping. Compare the results of experiments with the standard database. The correct rate can reach 70%. By this system, patients can sleep at home to monitor the sleeping quality and also evaluate the initial sleep disorders of patients. This can help to simplify the process or further treatment.