感知异构车载网络中多维资源协同优化和高效数据共享

The recent emergence of a broad range of vehicle-related cyber-services has stimulated the explosive demand growth of vehicle-based wireless networks to achieve high reliability, low delay, low energy consumption, and ubiquitous coverage. In particular, the rapid development of novel technologies in intelligent automobiles, autonomous driving, and smart traffic sensing can take advantage of the multi-domain capabilities of vehicular networks for information sharing, content caching, and cooperation in order to achieve highly efficient data networking among various nodes with diversified needs. These transformative applications and scenarios present many new challenges and opportunities for further improving road safety, traffic efficiency, and user experience. In response to these considerations, this research project centers on the innovative optimization of multi-domain resources and high efficient data sharing in multi-tier heterogeneous vehicle networks. The unique contribution and characteristics of this project stem from the methodology for integrative exploitation of efficient spectrum reuse, multi-node cache coordination, and aerial-ground wireless coverage. By jointly taking into account spectrum domain, coding domain, and spatial domain, this project team will innovatively utilize mathematical tools and analytic methods such as combinational optimization, matching theory, contract theory, and deep learning in order to develop novel methodologies, mechanisms, and schemes for overcoming the resource management obstacles that currently limit the widespread deployment of vehicle-based networks. The project specifically focuses on: 1) LTE-V based resource management for coexistence of hierarchical data services; 2) LTE-U based capacity improvement enhancement with contract incentives; 3) Robust and efficient coding in cache-based efficient content distribution and sharing; 4) UAV-assisted cross-domain optimization of communication-caching resource coordination. The project outcomes will provide highly valuable insights and contribute substantially to the foundations of intelligent vehicle networks in the future.

多元化新兴车辆业务的出现激化了车载网络对于高可靠、低能耗、广覆盖通信需求的提升。智能化汽车和新型通信感知等新技术的迅猛发展，使得通过扩展异质车辆节点通信、缓存、感知等多维能力进行高效数据传输和共享成为可能，为提高车辆道路安全、交通效率以及用户体验等带来了新思路和新挑战。因此，本项目将依托高效频谱复用、多节点协同缓存、通信立体覆盖的理念，从频谱域、编码域以及空间域等三种扩维思路出发，有机结合匹配、契约、组合优化和机器学习等理论，重点研究感知异构车载网络中多维资源协同优化和高效数据共享，致力于提出一套突破车载网络能力瓶颈的新型机理与方法。包括：1）支持分级化业务并存的LTE-V资源管理；2）契约激励下基于LTE-U技术的网络容量提升；3）基于编码缓存的高效内容分发与共享；4）无人机协助下的通信-缓存跨域资源协同优化。该项目的开展将为未来智能化车载网络的应用和落地实施奠定基础。

本项目围绕感知异构车载网络开展研究，通过扩展车辆节点的通信、缓存、感知等多维能力以突破车辆间的通信瓶颈，提升整体的网络容量，为增强车辆安全、交通效率和用户体验提供有力的技术支撑。本项目从频谱域、编码域以及空间域的扩维思路出发，利用匹配理、契约论、组合优化和人工智能等方法，提出了一系列新型高效的解决方案和算法，实现了面向差异化业务需求的多维资源协同优化和高效数据共享。包括：.1）研究了基于LTE-V的资源分配机制，利用分级化的资源管理方案，设计了一系列高效的资源分配算法，在有限频谱、能耗等资源的约束下，提升用户体验，满足差异化的业务需求；.2）研究了LTE-U系统中的网络扩容方法，引入授权频谱和非授权频谱的链路聚合，提出了一系列基于契约论的节点激励机制，在解决用户信息不对称问题的同时，实现网络吞吐量和系统稳定性的有效提升；.3）研究了时变动态通信环境下的高效数据共享方案，基于编码缓存的技术特点对抗不可靠的无线信道，并将通信优化扩展至通信-缓存协同优化，提出了一系列高效的跨域资源联合优化方法，有效提升了网络的谱效和能效并增强数据传输的可靠性；.4）研究了无人机辅助下的多维资源协同优化方法，基于无人机灵活易部署的特点，设计了一系列面向不同业务场景下融合无人机部署、路径规划、通信匹配和内容缓存等多目标的联合优化算法，实现多目标需求下的多维资源跨域协同与互补增强。.项目执行期间，共发表论文37篇，申请发明专利15项，完成技术报告3份，搭建了一套通信模拟平台，培养青年教师2人、博士后3人、博士研究生3人、硕士研究生10人，顺利完成了任务书的各项指标要求。