摘要： 近年来，车联网作为智能交通系统的核心组成部分，凭借其卓越的车辆互联特性，为用户提供智能化、便捷化和安全化的车载服务。为了进一步优化车载服务的服务质量（QoS），移动边缘计算（MEC）被深度整合于车联网中，旨在为车辆提供地理位置较近的计算资源，降低任务处理延迟和能耗。然而，传统的MEC服务器部署主要依赖于地面基站（BS），这不仅导致高昂的部署成本，而且限制其覆盖范围，难以确保为所有车辆提供无间断服务。为了应对上述挑战，空地协同车联网作为一种新兴的技术方案应运而生。无人机（UAV）能够借助其视距链路的灵活性动态地协助路边单元（RSU），为车辆用户提供更为灵活的计算资源，进而保障车载服务的连续性和高效性。因此，本文提出了一种基于空地协同的动态车载边缘任务卸载方法（DVETOM）。该方法采用车-路-空架构，构建了车辆到RSU链路（V2R）和车辆到UAV链路（V2U）。本文针对车辆任务的本地执行、卸载至RSU执行和卸载至UAV执行三种模式分别构建传输模型和计算模型，并以最小化系统时延和能耗作为联合优化目标构建目标函数。DVETOM将任务卸载问题转化为马尔可夫决策过程（MDP），基于深度强化学习（DRL）的分布式深度确定性策略梯度算法（D4PG）优化任务卸载策略。最后，通过与五种基准实验进行对比，DVETOM在提升车辆用户QoS的同时，在降低系统时延方面优于现有方法3.45%~23.7%，在降低系统能耗方面优于现有方法5.8%~23.47%。综上所述，本文提出的DVETOM方法有效地优化了车联网中的车载边缘任务卸载，为车联网用户提供了更高效、更节能的服务解决方案，展现了其在智能交通系统领域的广泛应用潜力。

Abstract: n recent years, the Internet of Vehicles as a pivotal component of Intelligent Transportation Systems, has been leveraging its outstanding vehicle interconnectivity to provide users with intelligent, convenient, and safe in-vehicle services. To further optimize the Quality of Service (QoS) of these services, Mobile Edge Computing (MEC) has been deeply integrated into the Internet of Vehicles, aiming to provide geographically proximal computing resources for vehicles, thus reducing task processing latency and energy consumption. However, traditional MEC server deployment primarily relies on terrestrial Base Stations (BSs), resulting in high deployment costs and limited coverage, making it difficult to ensure uninterrupted services for all vehicles. To address these challenges, the emerging technology of air-ground collaborative IoV has emerged. Unmanned Aerial Vehicles (UAVs) can dynamically assist Road-Side Units (RSUs) using their flexibility in line-of-sight links, providing more flexible computing resources for vehicular users, thereby ensuring the continuity and efficiency of in-vehicle services. Therefore, this paper proposes a Dynamic Vehicular Edge Task Offloading Method (DVETOM) based on air-ground collaboration. This method adopts a vehicle-to-road-to-air architecture, establishing Vehicle-to-RSU (V2R) and Vehicle-to-UAV (V2U) links. Transmission and computation models are constructed for three modes: local execution of vehicular tasks, offloading tasks to RSU, and offloading tasks to UAV. The objective function is established with the joint optimization goal of minimizing system latency and energy consumption. DVETOM transforms the task offloading problem into a Markov Decision Process (MDP), and optimizes the task offloading strategy using the Distributed Deep Deterministic Policy Gradient (D4PG) algorithm based on Deep Reinforcement Learning (DRL). Finally, through comparison with four benchmark experiments, DVETOM outperforms existing methods by 3.45%~23.7% in reducing system latency and 5.8%~23.47% in reducing system energy consumption, while improving QoS for vehicular users. In conclusion, the DVETOM approach presented in this paper has effectively enhanced the offloading of vehicular edge computing tasks within the Internet of Vehicles. It offers users of the Internet of Vehicles a more efficient and energy-conserving solution, showcasing its extensive potential for application within the realm of intelligent transportation systems.