Relay Selection Algorithm For Vehicular Ad Hoc Networks

The information field is undergoing a new round of technological revolution fromthe Internet to the Internet of things (IoTs). Vehicular ad hoc networks (VANETs) havereceived considerable attentions from both academia and industries as an importantintersection of the two fields of strategic emerging industries, IoTs and intelligenttransportation. In VANET, the fast movement of cars and the limited coverage of accesspoint (AP) result in the disconnection between some of the vehicles and APs. To addressthis problem, relay vehicles (RVs) can be applied to forward data packets for sourcevehicles (SVs). In the case that multiple candidate RVs are available, how to jointlyconsider the characteristics of physical channel, the linking access delay due to messagecollision, the loading condition of RVs and select the optimal RVs to support usercommunication requirements and achieve the optimum of system performance becomesan important research topic in VANET.Considering the scenario that multiple SVs, multiple available RVs and selfishvehicles coexist in VANET, a multi-objective optimization based RV selection algorithmis proposed in this thesis which jointly considers the service needs of SVs, the channelcharacteristics of RVs, the available bandwidth and the message collision of channelcompetition. The utility function of both SVs and RVs is formulated and maximizedachieve the performance optimization of all the SVs and RVs. Applying ideal pointmethod, the optimal RV selection strategy can be obtained.For the VANET application scenario of multiple cooperative SVs and multiple RVs,we propose a Game theory based relay selection algorithm. By jointly considering theimpacts of various factors on the performance of relay selection, a cooperative Gamemodel of SVs and RVs is formulated and solved via bipartite graph matching algorithm(Kuhn-Munkras algorithm), corresponding to the optimal relay selection strategy.Focusing on two typical VANET services, i.e., delay sensitive services andthroughput sensitive services, this thesis proposes a cluster-based relay selectionalgorithm. A utility function based optimal cluster head selection and cluster switchingalgorithm is proposed.The optimal relay selection algortihms are proposed for VANET in the light ofspecific network scenarios and the service requirements of users. This research work can be used as a reference to further study of relay selection in VANET, and exhibits certaininnovation, theoretical value and practical significance.