CRII: CIF: Towards Self-Powered Heterogeneous Cellular Networks

在移动数据流量空前增长的推动下，蜂窝网络正在经历一场巨大的范式转变，从以覆盖为中心的高功率蜂窝基站(基站)的同质部署，转向更有机的容量驱动型部署，其中还包括各种类型的低功率基站，统称为小小区。为了保持部署灵活性并限制运营成本，这些基站非常需要有能力通过独立的能量收集模块为自己供电。然而，与能量采集相关的固有的不可靠性，再加上小蜂窝的不规则位置，使得量化这种“自供电”的异质蜂窝网络的可靠性具有挑战性。该项目旨在通过新的分析工具和指标为这些网络奠定基础。自供电基站具有额外的自回程能力，将产生一个“真正的无线”蜂窝网络，从而实现各种“即插即用”部署。该项目的进一步广泛影响将通过研究传播、教育、扩大学生的参与和行业合作来实现。该项目的首要目标是为自供电的异质网络建立基本的性能限制，该项目采用了一种跨学科的方法，使用了通信、信息论和随机几何的工具。该项目的第一个协同部分开发了一个新的综合模型，能够捕获自供电异质网络的关键特征，如基站能力的差异、部署中的不规则性以及能量水平的不确定性。强大的数学工具以随机几何和点过程理论为基础，为该模型提供了易处理的能力，从而允许对自供电的异类网络特有的新性能指标进行正式分析。例如，由于基站之间的负载耦合，当一个基站耗尽其能量时，它可能会引发“级联效应”，导致整个网络随之崩溃。第一套结果描述了这一行为，并确定了网络保持稳定的制度。一旦稳定性得到保证，第二组结果集中在这种新模式下的最终用户性能，其中幸存的基站并不总是保证有足够的能量来满足其所有负载，从而导致能源中断。最后，将这些指标与经典的体验质量指标(如下行链路速率)进行联合分析。

Driven by unprecedented increase in mobile data traffic, cellular networks are undergoing a huge paradigm shift from a coverage-centric homogeneous deployment of high-power cell towers (base stations) to a more organic capacity-driven deployment that additionally includes various types of low-power base stations, collectively called small cells. To maintain deployment flexibility and limit operational costs, it is highly desirable that these base stations have a capability to power themselves through self-contained energy harvesting modules. The inherent unreliability associated with energy harvesting, however, combined withthe irregular locations of the small cells, makes it challenging to quantify the reliability of such "self-powered" heterogeneous cellular networks. This project aims to lay the foundations of these networks through new analytical tools and metrics. Self-powered base stations with additional capability to self-backhaul will result in a "truly wireless" cellular network thereby enabling a variety of "drop and play" deployments. Further broader impact of this project will be through research dissemination, education, broadening participation of students, and industry collaboration. With the overarching goal of establishing fundamental performance limits for self-powered hetergeneous networks, this project adopts a cross-disciplinary approach involving tools from communications, information theory, and stochastic geometry. The first synergistic component of this project develops a new comprehensive model capable of capturing key characteristics of self-powered heterogeneous networks, such as the differences in the base station capabilities, irregularities in their deployments, and uncertainties in their energy levels. Powerful mathematical tools with foundations in stochastic geometry and point process theory lend tractability to this model, thereby allowing the formal analysis of new performance metrics unique to self-powered heterogeneous networks. For instance, due to the coupling of loads across base stations, when one base station drains out its energy, it may initiate a "cascade effect" taking the whole network down with it. The first suite of results characterizes this behavior and determines the regimes in which the network remains stable. Once stability is guaranteed, the second set of results focuses on the end-user performance in this new paradigm where the surviving base stations are not always guaranteed to have sufficient energy to serve all their load, thereby resulting in energy outages. Finally, these metrics are jointly analyzed with classical quality-of-experience metrics, such as downlink rate.