臺灣大學: 電信工程學研究所謝宏昀張智華Chang, Chih-HuaChih-HuaChang2013-03-272018-07-052013-03-272018-07-052012http://ntur.lib.ntu.edu.tw//handle/246246/252629物聯網的興起帶動了物聯通訊的需求，許多物聯網的應用往往需要物聯裝置進行大量的資料傳輸；為了處理這些大量傳輸資料所衍生的問題，在本論文中，我們提出利用物聯資料彼此之間的相關性，以提昇「整體資料本身」而非「個別物聯裝置」品質的觀點來設計相關通信技術。為了驗證此觀點，我們考慮一個資料收集的應用，眾多物聯裝置透過無線傳輸將收集的資料回傳至資料收集中心。由於無線傳輸的資源是有限的，因此我們所要解決的問題就是在有限的資源下如何有效地做資源分配使得此資料收集的應用達到最好的效果。在此架構下，我們首先考慮物聯裝置只能做獨立的資料編碼與壓縮，並且提出以收集資料中有用的訊息為目標的無線資源分配技術。由於求得最佳解的方法需要相當高的複雜度，因此我們將問題拆解為兩個子問題：選擇服務裝置的子問題以及對選定裝置做資源分配的子問題。透過模擬的結果，可以發現我們所提出的方法雖然不以服務最多的裝置數量或達到最大的資料流量為目標，但是卻可以從相關性資料中收集到最多有用的訊息。為了更深入探討不同的編碼能力對物聯通信的效益，因此我們接著考慮物聯裝置可以做分散式編碼壓縮以及透過聆聽其他裝置的資料做共同編碼壓縮。這兩種編碼技術皆可以消除個別裝置資料與其他裝置資料相關的部分，避免多餘的資料重複傳輸；而分散式編碼更可以在不需要彼此交換資訊的情況下達到相近的編碼效果。我們透過模擬結果比較不同的編碼方式，發現分散式編碼技術比獨立編碼可以壓縮更多資料，因此可收集到更多資料中的訊息，尤其當資料的相關性較高時效果更為明顯。這也進一步驗證本論文所提出的觀點：在物聯通信下，應該以提昇「整體資料本身」而非「個別物聯裝置」品質的觀點來設計相關通信技術。Many applications involving machine-to-machine (M2M) communications are characterized by the large amount of data to transport. To address the ``big data'' problem introduced by these M2M applications, we argue in this thesis that instead of focusing on serving individual machines with better quality, one should focus on solutions that can better serve the data itself. To substantiate, we consider the scenario of data gathering in a wide area with clustering, where the cluster head (CH) collects the correlated data of machines in each cluster and feedback to the data aggregator through wireless links. Since each cluster has limited radio resources, the problem arises as to how the resources can be effectively utilized for supporting such an M2M application. We start with a simple scenario of independent coding and transmission, and propose an approach that takes into consideration ``useful'' information content of the correlated data for resource allocation. The NP-hard problem is solved by separating the resource allocation sub-problem from the node selection problem. Numerical results show that although the aggregate data rate or the number of machines that can be supported is not maximized, the data collected at the CH exhibits significant quality gain for the target M2M scenario. To further explore the performance gain under different coding schemes, we then consider the problems of machines applying distributed source coding and dependent source coding via overhearing to reduce redundancy of correlated data. Simulation results indicate significant benefits of distributed source coding especially for highly correlated data, thus substantiating our argument regarding the move towards data-centric communication for M2M applications.1191965 bytesapplication/pdfen-US物聯通訊資料收集資料相關性資源分配消息理論分散式訊源編碼Machine-to-machine communications (M2M)Data gatheringData correlationResource allocationInformation theoryDistributed source codingthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/252629/1/ntu-101-R99942106-1.pdf