Trajectory Data Mining-Based Routing in DTN-Enabled Vehicular Ad Hoc Networks

A delay/disruption tolerant network (DTN) architecture where a “store-carry-forward”strategy is adopted for data transmissions can be utilized in vehicular ad hoc networks (VANETs). The key point of routing in DTN-enabled VANETs is to choose the best node and determine the best time to forward messages. Time-space graph models provide an idea of converting the dynamic routing problems into static ones in deterministic DTNs. But it is a challenge to predict vehicles' future positions in order to obtain the time-space graph. In this paper, to achieve the cost-efficient and reliable routing in DTN-enabled VANETs, a novel timeliness-aware trajectory data mining algorithm is proposed to predict nodes' future positions. A sparse time-space graph is then obtained, based on which, two routing heuristics are proposed. Simulation results demonstrate that our proposed routing algorithms ensure low cost and high reliability over time.