Task offloading and resource allocation for intersection scenarios in vehicular edge computing

Abstract

Vehicular edge computing (VEC) is a promising paradigm to relieve the overload on corresponding edge servers by utilising the idle resources of nearby vehicles. However, due to the high mobility of vehicles, the vehicles perhaps drive out of the communication range of user equipments (UEs) during task processing. Therefore, it is important to select the appropriate vehicles as service nodes. In this paper, we propose a task offloading and resource allocation scheme for UEs near the intersection. We first study the availability of vehicles according to characteristics movement of vehicles at the intersection. Then, considering the delay constraint of tasks, and the computing capacity of vehicles and the edge server, a double deep Q-network approach is adopted to obtain the optimal policy of task offloading and resource allocation. Simulation results show that the proposed scheme has better performance in improving the average utility of UEs and task success rate.