雾无线接入网络的传输模式选择和资源分配方法

Owing to the centralized cloud computing, the system performance is greatly improved in cloud radio access networks (C-RANs) compared with traditional wireless networks. However, C-RANs suffer from capacity constrained fronthaul and long latency. To reduce the end-to-end transmission latency and alleviate the fronthaul capacity constraints, the fog computing based radio access network (F-RAN) is proposed as an enhanced evolution of C-RANs to make full use of caching and processing capabilities of edge devices. Up to now, little research on this novel networking paradigm has been done. To overcome the key theoretical and technological issues that can hinder the development of F-RANs, derive the theoretical performance bound, and realize the coordinated optimization of multi-dimensional resources, this application would focus on the following aspects: 1) Propose adaptive transmission mode selection methods, establishing related models for analyzing network performance, and finding the main factors influencing the theoretical performance bound; 2) Research game theory based resource allocation methods and raising the network performance with the help of edge devices; 3) Investigate cache-aware cross layer resource allocation methods which take the different caching capabilities of edge devices into account and meanwhile balancing the system performance and computing complexity. The outcomes can well reflect the performance gain brought by F-RANs, provide key technological solutions, and contain independent core intellectual property rights, laying the theoretical and technological foundations of the standardization and commercialization of F-RANs.

云无线接入网络通过引入集中云计算来提升传统无线网络性能，但受限于去程链路容量和时延过长等瓶颈。为了减少端到端传输时延和减轻去程链路容量要求，利用网络边缘设备的缓存和处理能力，业界提出了雾无线接入网络作为对云无线接入网络的增强演进。雾无线接入网络作为一种新颖的组网机制，相关理论研究基本空白。为突破制约雾无线接入网络发展完善的关键理论技术问题，揭示雾无线接入网络理论性能限，确保多维资源协同优化，本申请拟研究：1)提出自适应的传输模式选择方法，建立相应的网络性能理论模型，揭示影响理论性能限的关键因素；2)研究基于博弈理论的资源分配优化方法，充分利用边缘分布特征提升组网性能；3)研究基于缓存感知的跨层资源分配优化方法，体征缓存的差异性，设计性能和复杂度折衷的资源分配优化方法。研究成果有助于揭示雾无线接入网络的性能增益，提供关键技术解决方案，掌握核心自主知识产权，为其标准应用提供理论支撑和技术指导。

雾无线接入网络（F-RAN）通过对边缘设备的充分利用，可适应动态的网络负载并降低对去程链路和BBU 池的负担。为了充分发挥F-RAN性能优势，适配未来差异化业务需求，项目组围绕F-RAN关键理论技术瓶颈问题展开了研究，揭示了雾无线接入网络理论性能限制，实现了面向不同业务需求的多维资源协同优化。具体而言：首先，针对F-RAN不同的接入方式对性能影响难以表征量化的挑战建立了相应的网络性能理论模型，进而揭示了影响理论性能限制的关键因素；其次，针对F-RAN集中分布混合场景下多维资源相互耦合，难以联合优化的难题，提出了基于博弈理论的用户模式选择策略及资源分配优化方法，充分利用边缘计算能力提升F-RAN性能；最后，针对F-RAN边缘缓存放置复杂度高的挑战，研究了基于缓存感知的跨层资源分配优化方法，实现了资源分配优化方法性能和复杂度的折中。通过本项目基金的资助，现已在国际、国内学术期刊和会议上发表论文9篇（其中SCI论文5篇，EI论文4篇），申请发明专利5项（3项已授权），输出成果向标准化组织提交标准提案或报告3项。培养青年骨干教师2名，博士后1人，博士/硕士15名。