在蜂巢網路的機器通訊網路下之能量效益群組傳輸

Abstract

在機器網路通訊中，群組傳輸系統有益於降低能量消耗、資源分配以及提升服務的品質。我們可以使用分群演算法將系統中的機器設備分群並找出適合的協調者與基地台溝通以實現群組傳輸系統。目前提出的演算法是根據複雜度較低的K-means分群演算法去做修改。不過在這些修改過的演算法中有些問題會讓分群的結果不盡理想。在這篇論文中，由於機器設備都是裝載電池的，所以我們主要關注的是系統的能量消耗並且提出改良的演算法來解決之前的問題並且降低系統的能量消耗。除此之外，我們還考慮了一種擁有群組傳輸也同時具備直接傳輸的複合型系統去增加能量效益。使用我們所提出的演算法，能有效的減少系統的能量消耗。另外，群組的數量在分群演算法之中一直都是個不容易解決的問題。而使用我們所提出的演算法，會在分群的結果當中得到一些有關於群組數量的資訊。利用這些資訊，我們可以在演算法執行結束之後對分群結果做一些調整得到群組的數量，並且這個結果能夠達到我們期許的效能。

In the machine to machine communication, group based system is use for energy saving, resource allocation and quality of service. To implement group based system, we group the machine devices together and communicate with base station by grouping algorithm and coordinator selection scheme. These modified algorithms are based on K-means, which is a low complexity algorithm. However, the modified algorithms have some exist-ing problems, these problems will make the algorithm converge with inefficient grouping result. In this thesis, since the machine devices are battery equipped, we focus on the up-link energy consumption of system and propose the improved algorithm to solve the problems. Furthermore, we also consider a hybrid system with direct transmission and group transmission to increase energy efficiency. With the improved algorithm and hybrid system, we can minimize the energy consumption of system efficiently. In addition, amount of groups is always a problem in the grouping algorithm. With our proposed algorithm, there is an information about the amount of groups in the grouping result. We use this information to find out the amount of groups at the end of process and also maintain the performance of algorithm.