基于按需深度覆盖与D2D中继的毫米波网络容量优化研究

New applications for mobile Internet and ubiquitous interconnection for all things will generate a huge amount of data, which expects to be addressed in future mobile communication networks with extremely high capacity. An ultra-dense deployment of millimeter wave base stations can satisfy the high capacity requirement of future mobile communication networks, but the deploying cost of such a solution is extremely high and there is a low utilization rate for equipment. This project intends to build an economically viable and on-demand deep coverage network, which can reduce the number of base stations while meeting capacity requirement..The intended studies are as follows: 1) the technologies for depth coverage on demand, including virtual access point election, overlay network architecture, and depth coverage quality optimization; 2) the network capacity optimization schemes based on D2D relay mode under the non-C-RAN framework enhanced by depth coverage, including D2D relay selection, channel reuse, power regulation, and load balancing; 3) the network capacity optimization schemes based on traffic convergence of virtual access point and D2D relay bridge under a C-RAN framework, including establishment of clustering principle for intelligent objects, initialization distribution of virtual access points served as sinks, clustering algorithm design, and selection for D2D relay bridge; 4) the corresponding incentive mechanisms for stimulating wireless terminals to participate in virtual access point election and D2D relay service.

新型移动互联网应用与万物泛在互连所产生的海量数据传输需求，对未来移动通信网提出了极高的容量要求。超密集部署毫米波基站能满足未来移动通信的容量要求，但是部署成本极高，且设备利用率低。本项目拟利用具备毫米波通信接口的无线终端，基于博弈的激励机制，采用D2D中继方法，构建经济可行的按需深度覆盖网络，减少基站数量，同时满足容量要求。具体拟研究1）按需深度覆盖技术，包括虚拟接入点选举、覆盖网架构搭建、深度覆盖质量优化机制等；2）按需深度覆盖增强的非C-RAN构架下基于D2D中继模式的网络容量优化方案，包括D2D中继选择、信道复用、功率调节、负载均衡等；3）C-RAN构架下基于虚拟接入点汇聚与D2D中继桥接的网络容量优化方案，包括智能物端聚类原则与规律、作为汇聚点的虚拟接入点分布规律、新型聚类算法设计、D2D中继桥接点选择机制等；4）激励无线终端参与虚拟接入点选举与D2D中继服务的激励机制。

新型移动互联网应用与万物泛在互连产生了海量数据传输需求，但满足该需求的移动通信网络的构建成本高且设备利用率低。本项目基于按需深度覆盖思路和采用D2D中继方法研究了毫米波网络容量优化问题，为上述挑战性问题提供了一种有效解决手段。通过系统分析网络覆盖、容量优化、激励机制等问题的特点，提出了针对按需深度覆盖及激励问题、基于D2D中继的毫米波网络性能优化问题、基于机器学习的新型毫米波网络性能优化问题、促进蜂窝终端协作的激励问题的一系列解决方案。提出的新颖层次化智能认知架构方便了mmWave C-RAN部署和实现了超高容量；提出的基于BFT信息辅助的无线接入资源管理算法在平均能效和链路吞吐量方面都取得了更好的性能；提出的CRA-D2D算法和DRA-D2D算法在网络能量效率和网络容量方面均优于现有算法；提出的MHRA算法在保证网络容量和收敛速度在可接受水平的同时，在网络能量效率方面明显优于同类算法；提出的基于Q学习机制的传输容量稳定方案可以实现更稳定的传输能力和更高的能量效率；提出的基于在线学习的毫米波波束选择机制能够显著提升蜂窝辅助车载系统的网络容量；提出的促进终端协作提升蜂窝覆盖质量的激励机制能够显著提升蜂窝弱覆盖区的覆盖质量；提出的促进终端协作提升蜂窝边缘设备频谱效率的激励机制能够显著提升蜂窝边缘设备的频谱资源利用率；提出的促进终端协作提升能量受限智能感知设备寿命的激励机制能够有效节省蜂窝边缘智能感知设备的能量。