Vehicle-to-Vehicle (V2V) communication can not only provide unrestricted inter-vehicle information transmission, but also improve spectrum utilization efficiency. However, it also brings uncontrollable co-channel interference, which can not guarantee the quality of service of V2V communication. In this paper, we propose an intelligent resource allocation scheme for V2V communication to improve vehicle connectivity. To enhance cooperation among vehicles and avoid excessive co-channel interference between them, we propose an asynchronous resource allocation method where vehicles choose to send or not to send data based on observed environmental information to ensure stable overall performance. Furthermore, we present a novel resource allocation algorithm based on Heterogeneous Agent Proximal Policy Optimization (HAPPO) to solve the resource allocation problem in asynchronous vehicular networks. The HAPPO algorithm calculates the global advantage function when each agent makes an action during the training process to ensure that the action taken contributes to the overall performance improvement. Our proposed approach improves the robustness of V2V communication by reducing co-channel interference while maintaining stable overall performance. Simulation results show that the proposed approach can effectively improve the V2V communication connectivity and reduce the packet loss rate compared with the existing methods.

中文翻译：

---

通过基于异构智能体近端策略优化的资源分配算法提升车辆连接性


车对车 （V2V） 通信不仅可以提供不受限制的车间信息传输，还可以提高频谱利用效率。但是，它也带来了不可控的同频干扰，无法保证 V2V 通信的服务质量。在本文中，我们提出了一种用于 V2V 通信的智能资源分配方案，以提高车辆的连通性。为了加强车辆之间的协作，避免车辆之间过度的同频干扰，我们提出了一种异步资源分配方法，即车辆根据观测到的环境信息选择发送或不发送数据，以保证整体性能的稳定。此外，我们提出了一种基于异构代理近端策略优化 （HAPPO） 的新型资源分配算法，以解决异步车辆网络中的资源分配问题。HAPPO 算法计算每个智能体在训练过程中做出一个动作时的全局优势函数，以确保所采取的动作有助于整体性能提升。我们提出的方法通过减少同信道干扰，同时保持稳定的整体性能，从而提高了 V2V 通信的鲁棒性。仿真结果表明，与现有方法相比，所提方法能够有效提高 V2V 通信连通性，降低丢包率。