Data Forwarding at Intersections in Vehicular Social Networks

Abstract

In vehicular social networks, the vehicles in different communities have different data requirements and different mobility models. Traditional researches focus on the forwarder selection in vehicle-to-vehicle communication, and hence the frequent forwarder change affects the overall performance. In vehicular scenarios with roadside units deployed at intersections, how to select an appropriate forwarding direction based on the social features is a key issue. In this paper, the roadside unit gathers the social attributes of vehicles on the roads by a well-designed information collection method, in which the odd-hop nodes and the even-hop nodes take different tasks to reduce resource cost. Furthermore, the direct forwarding contribution ratio is computed according to the social attributes, and together with the connectivity and the empirical direction selection of similar data, the forwarding priority of each candidate direction is calculated. Finally, the forwarding directions are selected and the replicas of data packets are distributed in these directions. Experiments with real road map and three communities show that, compared with traditional schemes, the proposed scheme has a high delivery ratio and a small transmission overhead (the number of V2V and V2I data transmissions) while maintaining an acceptable delay.