动态多信道无线网络中节点信道交汇及网络容量最大化问题研究

Dynamic multi-channel wireless networks such as Cognitive Radio Networks(CRNs) are key approaches to improve spectrum efficiency, which are currently the focus of study on wireless network models. However, among existing research results, for rendezvous problem (how the nodes choose the same available channel at the same time slot to establish a connection), it was generally ignored that channel environment would change over time; for network capacity maximization problem (how to make as many links as possible transfer data successfully in a time slot), rendezvous among multiple links was not explicitly considered and the influence of the rendezvous strategy to network capacity was not analyzed either. In this proposal, taking CRNs as background, we will derive efficient rendezvous strategies for two nodes under the channel environment changing with both time and space; with rendezvous explicitly considered, in network capacity maximization problem, we are to investigate strategies that make nearby links tend to rendezvous on different working channels to avoid interference. We plan to integrate rendezvous and link scheduling to design distributed approximation algorithms to maximize network capacity. Our research not only has great value in the application of dynamic multi-channel wireless networks, such as Cognitive Radio Networks, Wireless Sensor Networks and Internet of things, but also has theoretical value to enrich the study of distributed approximation algorithms under dynamic environments.

认知无线电等动态多信道无线网络是提高频谱利用率的关键手段，是当前重点研究的无线网络模型。然而，在动态多信道无线网络现有成果中，节点信道交汇问题的研究（节点如何通过同时选择同一公共可用信道而实现连通）普遍忽略信道环境随时间的变化；网络容量最大化问题的研究（如何使尽可能多的链路在同一时隙成功传输）普遍没有显性考虑实现多链路的节点信道交汇的方法及其对网络容量的影响。本项目将以认知无线电网络为应用背景，在随空间、时间都动态变化的信道环境下，为单链路上两个节点设计快速交汇策略；在显性考虑多链路信道交汇的前提下，设计策略使邻近链路倾向于交汇在不同的信道上以避免干扰，综合交汇策略与链路调度设计分布式近似算法最大化网络容量。本项目研究成果不但在诸如认知无线电网络、无线传感器网络、物联网等存在动态多信道环境的网络中具有应用价值，也对丰富动态环境下的分布式近似算法研究具有理论价值。

项目主要研究认知无线电网络、移动计算领域无线信道分配与调度等网络性能优化关键问题。具体研究了动态多信道交汇问题的相关理论，网络资源（如信道）拥塞控制博弈， 混合网络中的信道在线调度和网络吞吐量优化，以及移动计算领域（如边缘计算）无线传输的在线调度和管理。在相关问题中提出了一系列有理论保证的算法解决方案，并通过实际数据或系统进行了验证。项目相关研究迄今共发表论文12篇，包括领域内顶级论文CCF A类6篇，B类2篇。所发表的成果得到领域内专家的关注，单篇论文发表一年内引用最高达30余次（Google数据），其中包括了12个来自Princeton，UIUC等著名大学的ACM/IEEE Fellow团队。项目执行期间搭建了包括10余台USRP，可编程交换机在内的无线网络实验平台，培养了多名研究生及本科生。