Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/54400
Title: | 適用於無線照護系統之個人化基底壓縮感知還原演算法 Compressive Sensing Reconstruction Algorithm with Personalized Basis for Wireless Healthcare System |
Authors: | Yi Chen 陳奕 |
Advisor: | 吳安宇(An-Yeu Wu) |
Keyword: | 心電遠距監控,壓縮感知,稀疏信號重建,字典學習, ECG telemonitoring,compressive sensing,sparse signal reconstruction,dictionary learning, |
Publication Year : | 2015 |
Degree: | 碩士 |
Abstract: | 在高齡化社會以及慢性病增長的趨勢下,居家照護的需求大增。其中如何整合可攜式感測器與無線通訊來實現無線照護系統,正是實現居家照護的重點方向。在無線照護系統的訊號處理方面,系統須連續不斷的偵測各種生理訊號以提供即時的病情監控,而眾多的訊號將大量消耗系統的頻寬以及功耗。面對如此困境,我們提出利用目前最新的訊號處理技術-壓縮感知(compressive sensing)來解決。壓縮感知可以將高維度的稀疏訊號,透過測量矩陣取得低維度的測量值,因此系統只需要以低維度的訊號做傳遞,等需要時再利用範數極小化等方法將低維度的取樣重建回高維的訊號。利用壓縮感知,我們能對無線照護系統提供更有效率的訊息傳遞方式。
壓縮感知的基礎建立在信號的稀疏特性上,信號必須足夠稀疏我們才有機會將其還原。然而,傳統事先建好的基底並不能很好的讓生理信號變得稀疏,在稀疏性不夠的情況下,我們將無法還原回原本的信號。在這篇論文中,我們利用字典學習的技術為每個病人建立他的個人化基底(personalized basis)來幫助還原。利用個人化基底,我們可以為壓縮感知無線照護系統提供更低的稀疏度跟更好的壓縮效果。此外,考慮到生理信號隨時間的變異性,我們提出了生理變異偵測的技術與低複雜度字典更新演算法做結合,使系統在信號特性有改變時能維持好的壓縮效果。在信號重建方面,我們將可適性濾波器的演算法引進到壓縮感知的還原問題中,提出一個低複雜度的隨機梯度追蹤(stochastic gradient pursuit)還原演算法。總結來說,提出的壓縮感知無線照護系統能達到更好的壓縮率以及更低的資料傳輸功耗。 With the aging of society and rising of chronic diseases, the demand for wireless home care has been increasing substantially. In the signal processing aspect of wireless healthcare system, the system needs to detect various physiological signals continuously and provides real-time condition monitoring. These numerous signals consume large bandwidth and power in the system. Faced with such dilemma, we propose to exploit the newest signal processing technique, compressive sensing, to resolve these problems. Compressive sensing obtains low dimensional measurements by sampling on high dimensional sparse signals with measurement matrix. Thus, the system only transmits low dimensional signals, while the original high dimensional signals can be recovered by norm-minimization method. With compressive sensing, we can provide a more efficient way to transmit information in wireless healthcare system. Compressive sensing is based on the sparse feature of signal. The signal needs to be sparse enough for recovery. However, common predefined basis cannot sparsify the biosignals well, resulting in the failure of reconstruction. In this work, we apply dictionary learning to construct the personalized basis for every patient in compressive sensing reconstruction. The personalized basis has sparser representation than common predefined basis. With the personalized basis, the system can achieve lower sparsity and better compression ratio. Moreover, considering the change of signal characteristic, we propose the physiological variation detection technique collaborated with the proposed low-complexity dictionary refreshing algorithm to maintain the good compression ratio. In terms of signal reconstruction, we introduce adaptive filters to the reconstruction problem in compressive sensing. A low-complexity stochastic gradient pursuit reconstruction algorithm is proposed. In conclusion, the proposed compressive sensing based wireless healthcare system can achieve higher compression ratio and lower data transmission power. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/54400 |
Fulltext Rights: | 有償授權 |
Appears in Collections: | 電子工程學研究所 |
Files in This Item:
File | Size | Format | |
---|---|---|---|
ntu-104-1.pdf Restricted Access | 3.91 MB | Adobe PDF |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.