Federated Learning Based Resource Allocation in Internet of Vehicles

车联网（IoT）是从智能交通系统衍生的网络，其允许道路车辆通过无线自组织网络交换车辆信息。高性能的车联网汽车可以大大提高道路交通效率，减少事故，实现高质量的信息娱乐系统等。然而，资源（例如，支持高性能车联网的物联网（网络、能源）是有限的，这对高性能车联网构成了严峻挑战。与此同时，车联网收集和交换的信息量是巨大的，比如每秒1GB。因此，本博士研究将重点关注车联网所需各种资源的高级分配和利用策略，考虑车辆移动性的影响，同时保持服务质量。该项目将研究广泛的现代技术，包括机器学习，能量收集，无人机，卫星网络等。最大化车联网的资源分配和利用。已经计划了几个工作包来进行一组理论优化，有效的算法/协议和全面的实验评估的研究。我们的目标是为新兴的车联网应用做出有益的贡献，使这些应用更加环保，并提高用户的驾驶体验。

Internet of Vehicles (IoVs) is a network derived from an intelligent transportation system, which allows on-road vehicles to exchange vehicle information through a wireless ad hoc network. High-performance IoVs can massively contribute to road traffic efficiency, reduce accidents, enable quality infotainment systems, and more. However, resources (e.g., networking, energy) supporting high-performance IoVs are limited, forming a critical challenge for high-performance IoVs. Meanwhile, the amount of information collected and exchanged by IoVs is massive, say 1GB per second. This PhD research will hence focus on advanced allocation and utilisation strategies for various resources required by IoVs, considering the impact of vehicle mobility while maintaining the quality of services.The project will investigate a wide range of modern technologies, including machine learning, energy harvesting, unmanned aerial vehicles, satellite networks, etc., to maximise resource allocations and utilisation for IoVs. Several work packages have been planned to carry out the study of a set of theoretical optimisations, efficient algorithms/protocols, and comprehensive experimental evaluations. The aims are to make useful contributions to emerging IoV applications so that such applications will be more environmentally friendly and enhancing of users' driving experience.