應用於生醫方面的CMOS全整合智慧型系統之設計及實現

隨著人口高齡化的持續惡化，健康照護成為了近年來最受關切的課題之一，尤其是針對老年人的照護。CMOS全整合系統仍然具有很大的發展潛力，並且將在未來於建立對健康照護友善的環境中扮演一個不可或缺的角色。本論文將對於使用系統單晶片(SoC) 技術的CMOS全整合智慧型系統做一個全面性的介紹。關鍵子電路和感測器的設計概念及實體操作展示也都有做描述。除此之外，三個針對不同醫療電子應用所特別設計，使用不同CMOS系統單晶片及感測器的全整合智慧型系統的設計內容也都會一一揭露。以下為所研究之應用於生醫方面的全整合智慧型系統的摘要介紹： 1. 以雙模式脈衝式射頻電刺激來即時舒解三叉神經痛的可植入式系統單晶片此系統單晶片是利用TSMC 0.35μm CMOS製程技術來實現，所針對的病症是三叉神經痛。三叉神經痛是源自於三叉神經發炎所引起在面部的極度疼痛感及肌肉痙攣。馬上抑制三叉神經痛發作時所引起的劇烈疼痛是其解決方案的第一目標。此使用脈衝式射頻電刺激的可植入式系統只需要簡單地透過手持裝置的控制，就可以隨時隨地以電刺激的方式來舒緩三叉神經痛。利用雙模式的脈衝式射頻電刺激以及可重組式的刺激參數可以針對每個不同的病患使電刺激的療效達到最佳化。溫度感測與訊號的超取樣可進一步的確保此系統使用上的安全性。在即時電刺激參數的監測及動物實驗的部分都有進行實體操作來對此系統的功能做驗證。 2. 應用於睡眠呼吸中止症治療之具無線監測及充電功能的智慧型牙套此智慧型牙套中使用的系統單晶片是利用UMC 0.18μm CMOS製程技術來實現，所針對的病症是睡眠呼吸中止症。傳統的牙套療法對於治療睡眠呼吸中止症來說是十分有效果且受歡迎的，但是針對個別病患來達到最佳的療效是非常耗時間的。此智慧型牙套的設計具有自動收集在治療過程中關於舌頭位置移動的生理資訊之功能。為了偵測來自於舌頭的微小壓力，一個具有超高靈敏度的穿隧式感測器陣列也整合在此系統之中。一個尺寸為 4.5×2.5×0.9 cm3的系統原型模組整合在樣本牙套中用來做為實作展示在連續時間中偵測舌頭位置移動的功能，此功能在量測的結果中得到了驗證。此系統設計的目的是增加牙套療法的效率，並成為解決睡眠呼吸中止症的完整解決方案。 3. 以雲端計算輔助之無氣囊可穿載式脈壓即時監測整合系統此系統所使用的晶片是利用UMC 0.18μm CMOS製程技術來實現，所針對的應用是即時脈壓監測。脈壓的變動跟動脈中的血流變化是息息相關的，其是在心血管疾病的早期診斷中的一個重要指標。為了偵測來自於腕動脈脈搏的微小壓力，一個具有超高靈敏度的穿隧式感測器也整合在此系統之中。在讀出電路將來自脈壓變化的訊號放大並轉換為數位訊號之後，系統的無線通訊模組將此資料以無線傳輸到智慧型裝置及雲端儲存設備中。即時並持續的脈壓監測從感測器端的偵測到最後的雲端計算，此完整的系統功能都在此做實作的展示。透過數位訊號的處理，長期的脈壓監測可以改善無氣囊裝置的準確度，而量測到的數據可以透過雲端的多種運算法來完成心血管突發事件的快篩。

關鍵字

系統單晶片；脈衝式射頻電刺激；三叉神經痛；牙套；穿隧式感測器；睡眠呼吸中止症；無氣囊

並列摘要

As the population aging keeps growing, healthcare is one of the most concerned topic in recent years, especially for the elderly. CMOS fully-integrated system is going to play an indispensable role in building a healthcare-friendly environment and still has great potential for further development. An overview of CMOS fully-integrated smart biomedical systems using the system-on-a-chip (SoC) technique is presented in this dissertation. Design concepts of key circuit blocks, sensors, and demonstration are also addressed. Three smart systems incorporated with different CMOS SoCs and sensors, which are specifically designed and implemented for different biomedical applications, are totally put forth. First, a dual-mode pulsed radio-frequency (PRF) stimulation SoC fabricated in TSMC 0.35μm CMOS process is realized for trigeminal neuralgia alleviation on demand. Trigeminal neuralgia is an inflammation of the trigeminal nerve, causing extreme pain and muscle spasms in the face. To stop this extreme pain immediately when it occurs is the top requirement for a solution to trigeminal neuralgia. The implantable system using PRF stimulation provides prompt pain alleviation by simply controlling a handheld device anywhere anytime. Dual-mode PRF stimulation and reconfigurable parameters are utilized to optimize the effectiveness for each patient. Temperature sensing and signal oversampling are also employed to ensure the safety. Real-time stimulation parameter monitoring and animal studies are demonstrated for the verification of this system. Next, a wireless monitoring smart oral appliance with a tunneling sensor array for sleep apnea treatment is presented. Conventional oral appliance therapy is effective and popular for OSA treatment, but making a perfect fit for each patient is time-consuming. This smart oral appliance, which is capable of intelligently collecting the physiological data about tongue movement through the whole therapy, is specifically designed. A tunneling sensor array with an ultra-high sensitivity is incorporated to accurately detect the subtle pressure from the tongue. A compact prototype module, whose size is 4.5×2.5×0.9 cm3, is implemented and embedded inside the oral appliance to demonstrate the tongue movement detection in continuous time frames. The functions of this design are verified by the presented measurement results. This design aims to increase efficiency and make it a total solution for OSA treatment. Last, a cuffless wearable system with cloud computing assistance is realized for real-time pulse pressure monitoring. The pulse pressure variation associated with arterial blood flow is a valuable indicator for early diagnosis of cardiovascular diseases. To detect the subtle wrist artery pressure, a tunneling sensor array with an ultra-high sensitivity is utilized in this system. After the readout circuit amplifies and converts the pulse pressure-induced signal, a wireless communication module transmits the data to intellectual devices and eventually to the cloud storage. Real-time and sustainable pulse pressure monitoring is demonstrated from sensor detection to cloud computing. Through the digital signal processing, long-term pulse pressure monitoring improves the accuracy of cuffless applications, and measured data can accomplish cardiovascular event fast screening after being processed by Hilbert Huang transform (HHT), pulse rate variability (PRV), and multi-scale entropy (MSE) analysis.