Socially-Aware Trajectory-based Routing in Vehicle Social Network

• Main Authors: Chang, Ching-Ju, 章晴茹
• Other Authors: Chang, Ing-Chau
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/gws9nz

碩士 === 國立彰化師範大學 === 資訊工程學系 === 105 === In recent years, as the rising of social networking, many vehicular packet forwarding protocols with the social concepts are proposed. The socially-aware vehicular forwarding protocols choose the appropriate forwarding vehicle by calculating values of the social relationship between the vehicle and the packet destination. However, these protocols do not consider the influences of different time period to the social relationship and the trajectory information of vehicles altogether. Hence, these protocols need to be improved. In this thesis, we propose a Socially-Aware Trajectory-based Routing (SATR) protocol which combines the value of social relationship of different time period into the trajectory-based routing protocol proposed before to improve the performance of packet delivery in Vehicular Social Network (VSN). It offers the following contributions: (1) Propose a new packet forwarding protocol in Vehicular Social Network to improve the disadvantages of lacking the use of time impact in conventional vehicular social network protocols. (2) Two major parts of SATR, i.e., the trajectory prediction stage and the packet forwarding one, are proposed. (3) In the trajectory prediction stage, we proposed algorithms to analyze the vehicle trajectories and social relationship on different time periods, design a new vehicle encounter sequence graph and construct the packet forwarding graph to reduce the time complexity of this stage. (4) In the packet forwarding stages, we propose the packet delivery algorithm and the exception handling one to improve the packet delivery ratio. (5) We execute simulation for related work and SATR by the well-known VANET simulator, the ONE (The Opportunistic Network Environment simulator), under different values of parameters. These results exhibit SATR outperform traditional routing in VSN. In addition, SATR uses fewer simulation times and memory space than Trajectory-based Routing and only about 5% to 10% lower than Trajectory-based Routing in packet delivery raio and delay.