Reliable routing and efficient message delivery in vehicular ad-hoc networks (VANETs) is a significant challenge owing to underlying environment constraints, such as dynamic nature, mobility, and limited connectivity. With the increasing number of machine learning (ML) applications in wireless networks, VANETs can benefit from these data-driven predictions. In this letter, we innovate and investigate ML-based classifications in VANETs to predict the most suitable path with the longest compatibility time and trust using a fog node based VANET architecture. The proposed scheme in SUMO VANET traces achieves up to a 16% packet delivery ratio (PDR) with a 99% accuracy and longer connectivity with only 3 ~ 4 hops, compared with existing AOMDV and TCSR solutions with merely a 4% PDR.

中文翻译：

---

DT-VAR：基于决策树预测兼容性的车载 Ad-hoc 可靠路由

由于潜在的环境约束，例如动态特性、移动性和有限的连接性，车辆自组织网络 (VANET) 中的可靠路由和高效消息传递是一项重大挑战。随着无线网络中机器学习 (ML) 应用程序数量的增加，VANET 可以从这些数据驱动的预测中受益。在这封信中，我们创新并研究了 VANET 中基于 ML 的分类，以使用基于雾节点的 VANET 架构来预测具有最长兼容性时间和信任度的最合适路径。与仅具有 4% PDR 的现有 AOMDV 和 TCSR 解决方案相比，SUMO VANET 跟踪中提出的方案实现了高达 16% 的数据包交付率 (PDR)，具有 99% 的准确度和更长的连接性，仅 3~4 跳。