A V2X-Enabled Framework for Delay-Sensitive Task Offloading in LAPS-Aided Internet of Vehicles

This paper explores the applications of mobile edge computing (MEC) and low altitude platform station (LAPS) in the vehicle-to-everything (V2X)-enabled Internet of Vehicles (IoV). In the LAPS-assisted IoV, vehicular applications consist of multiple interdependent subtasks, and vehicles can offload some subtasks to LAPS for further computation to alleviate the computing resources shortage at vehicles. To minimize the computing time, we propose a novel offloading scheme and formulate a comprehensive task latency minimization problem that optimizes offloading decision and resource allocation, taking into account subtask dependencies, imperfect channel state information (CSI), and energy consumption. Given the non-convex of the problem, we separate it into task offloading subproblem and resource allocation subproblem and design a novel algorithmic framework to determine the optimal task offloading and collaborative communication strategy. First, we derive the optimal spectrum reuse scheme and accordingly adjust the transmission power of vehicles adopting the proposed optimal resource allocation algorithm. Then, we introduce a task offloading algorithm that considers directed acyclic graphs (DAGs) to ensure that the dependencies among subtasks are maintained. Simulation results show that our strategy significantly reduces task processing latency and improves the quality of service (QoS) for vehicle users compared to existing approaches.