Federated Offloading Scheme to Minimize Latency in MEC-Enabled Vehicular Networks

Abstract

The vehicular networks with mobile edge computing (MEC) provide a promising paradigm to meet the explosive vehicular computing demands. To further reduce the total latency and improve the utilization of computation resources, we consider the available vehicular resources and propose the federated offloading of vehicle-to-infrastructure (V2I) and vehicle-to-vehicle (V2V) communication in MEC-enabled vehicular networks. In this paper, we investigate the problem of how to realize effective federated offloading for the moving vehicles, with the target to minimize the total latency. The computation task is divided into three parts: the part to compute locally, the part to offload to the MEC server (in the roadside) through V2I communication, and the part to offload to the neighboring qualified vehicles through V2V communication. Then, we propose two federated offloading modes according to the offloading order of V2I and V2V to find the optimal total latency results, considering the task allocation ratio among the three parts, as well as communication and computation environment conditions. Moreover, in order to use the available resources in the neighboring vehicles, we propose a distributed algorithm to obtain an optimal routing to offload the task of V2V part. Simulation results show that our proposed scheme can enhance the utilization of computation resources and decrease the latency.