A Vehicle-Assisted Data Offloading in Mobile Edge Computing Enabled Vehicular Networks

Abstract

With the emerging applications of vehicular networks, how to provide sufficient communication and computation supports are the two most important challenges for vehicular communication systems. Cloud-based vehicular networks and mobile edge computing frameworks have been proposed to relieve the computing burden of vehicles. However, for time-sensitive and computation-intensive applications with large input data size, e.g. image aided navigation, the data transmission process occupies large bandwidth, which may degrade the quality of service of all network users, especially in high density scenarios. Thus, in this paper we formulate a computation offloading problem for these time-sensitive and computation-intensive applications to minimize the data transmitted to the server. We propose two approaches to solve the formulated problem, i.e. a graph theory based method and a heuristic algorithm. Simulation results demonstrate that both algorithms can achieve near optimal solutions and greatly reduce the data volume transmitted to the server.