Routing Policy for D2D Group Communication

碩士 === 國立中山大學 === 資訊工程學系研究所 === 106 === Cisco predicted that mobile traffic will reach over 120 exabytes per month in 2018. Because of that, the Third Generation Partnership Project (3GPP) proposed a technology called “Device-to-Device (D2D)” communication. The first mention of this technology is in...

Full description

Bibliographic Details
Main Authors: Chao-ping Yang, 楊超評
Other Authors: Wei-Kuang Lai
Format: Others
Language:zh-TW
Published: 2017
Online Access:http://ndltd.ncl.edu.tw/handle/b4y59r
Description
Summary:碩士 === 國立中山大學 === 資訊工程學系研究所 === 106 === Cisco predicted that mobile traffic will reach over 120 exabytes per month in 2018. Because of that, the Third Generation Partnership Project (3GPP) proposed a technology called “Device-to-Device (D2D)” communication. The first mention of this technology is in 3GPP release12. D2D communications is a novel technology allowing user exchanges of their data over direct links instead of through base stations. D2D can improve system capacity, system throughput and data rate. It also can provide lower latency for user. Moreover, D2D communications play an important role in traffic offloading. In recent years, there are serval research indicating that people will construct stable social structures when they are geographically close. People would like to help each other when they are in the same group. It means that they would help relay data and cache data for other group members. In this paper, we designed an utility function, which consider delay and throughput at the same time, to evaluate the path. Because of the unit of delay and throughput are different, we defined how to standardize them. We also combine the advantages of Bidirectional search and Lifelong Planning A*, propose a new routing policy for group members, which can meet their demands with lower latency and better throughput. Furthermore, we device an utility function in our algorithm to select paths. Our simulation results have shown that the proposed mechanism can choose the path which has better throughput and lower latency. In addition, the proposed mechanism can apply to dynamic topology efficiently.