面向5G的超密集异构网络能效优化理论与方法研究

To support the rapid growth of mobile service, the dense hierarchal heterogeneous cellular networks will be utilized in the fifth generation mobile communication systems. However, the exponential growth in demand of mobile service will lead to more energy consumption. Therefore, how to improve the energy efficiency will be a great challenge for 5G. This project will investigate the mechanisms and methods of energy efficiency optimization in the hierarchal heterogeneous cellular network based on network deployment and radio resource management, including: (1) Based on 3-D Poisson Point Process, the distribution of communication nodes are modeled in ultra-dense scenarios. The new general interference and energy efficiency models will be proposed based on stochastic geometry theory. (2) In order to realize the elastic matching of base station resources and service demand, all base stations are configured dynamically according to the space-time characteristic of mobile service. Then the base station sleeping mechanism will be designed with inter-layer cooperation. (3) The inter-relationship among base station densities, transmit power and energy efficiency will be derived. In addition, we will build the energy efficiency optimization model in order to realize the joint optimization of base station density and transmit power. (4) We will establish the dynamic access mechanism with joint optimization of energy efficiency and network resources. Also, the relationship between energy efficient resource management and network environment will be revealed in this project. We will allocate the radio resource adaptively based on the attributes of network, users and service to improve the network energy efficiency. The achievements of this project will provide theoretical basis and technical support for the network deployment and radio resource management of 5G mobile communications.

为了支持移动业务的密集增长，5G移动通信系统将采用超密集的异构网络部署方式，同时业务的快速增长将带来能耗的剧增，因此如何提高系统能效将是5G面临的重大挑战。本项目将研究分层异构蜂窝网络的高能效机理和方法，实现网络部署与无线资源管理的能效优化，包括：(1)基于随机几何理论，准确描述分层异构网络中各类节点的空间分布，构建超密集场景下网络的立体分析模型；(2)基于业务的时-空特性动态地配置基站，通过层间协作设计基站休眠机制，实现基站资源与业务需求的弹性匹配；(3)推导基站密度及发射功率与能效的关系式，刻画它们的内在关系，构建能效优化模型以实现基站密度与功率的联合优化；(4)建立能效与资源联合优化的动态接入机制，揭示高能效资源管理与网络环境的关系，根据网络/用户/业务的多种属性自适应地分配资源，以提高网络能效。本项目的研究成果将为5G网络的部署及高能效无线资源管理策略提供很好的理论依据和技术支撑。

本项目针对超密集蜂窝异构网络的能效问题，基于随机几何理论进行干扰建模与分析，揭示了约束网络性能的因素，提出了超密集异构网络的高能效机理和方法。主要创新如下：（1）基于3-D随机几何理论，将超密集小基站立体分布抽象为3-D空间泊松点过程，构建了超密集场景下异构蜂窝网络立体分析模型，推导了动态功率控制准则下的覆盖概率和传输速率的闭合表达式；（2）提出了一种异构蜂窝网络的2-D PPP和3-D PPP混合干扰模型，分析了异构网中的干扰特性，推导了网络中目标用户的下行覆盖概率和传输速率；（3）构建了宏基站PPP分布和小基站Matern簇分布的混合模型，基于全概率公式，推导了网络的平均覆盖率和空间频谱效率的表达式；（4）基于3-D PPP推导了小基站蜂窝网络中用户ACR（Average Connection Ratio）的数学表达式，并对其进行单调性分析，获得了在ACR和平均速率双重约束下小基站的最优节能休眠策略，降低了网络能耗；（5）基于随机几何和排队论，构建了一种考虑动态业务的K层异构网络时空分析模型，推导了RR和RS两种调度策略时的平均吞吐量，分析了空间和时间随机性、偏置用户接入策略及频谱复用策略对网络平均吞吐量的影响，对接入与复用策略进行了优化，提高了系统容量；（6）针对多用户异构网络，提出了能量效率的子载波与功率联合优化算法，构建了基于能量效率最优的子载波配对、用户指派和功率分配联合优化模型，推导了等效信道增益，基于Dinkelbach方法将非线性混合整数规划问题转换为准凸规划问题，运用对偶理论和匈牙利算法求解联合优化问题，有效提高了系统的能效；（7）针对能量采集缓存辅助中继多用户协作网络，构建了数据包和能量包的耦合处理排队模型，运用拟生灭过程得到稳态概率，推导了系统吞吐量、时延和丢包率表达式，证明了基于APT（Adaptive Power Transmission）的数据与能量耦合机制可以有效改善系统性能。基于以上研究，本项目共发表论文65篇，其中SCI检索27篇，EI检索22篇，申请专利35项，其中12项已获授权，培养毕业博士生3名，硕士生25名, 获江苏省科学技术一等奖。本项目研究成果可以为超密集蜂窝异构网络的部署及高能效资源管理策略提供很好的理论依据和技术支撑。