CAREER: Data Dissemination in Multihop Wireless Networks: Theory and System Design

数据传播，即将数据或控制消息从一个端点传输到多个端点，是无线网络运营中不可或缺的组成部分。多跳无线网络中现有的和新兴的应用，如分布式工业控制、环境监测、应急响应系统，作用于无线设备的聚合，而不是单个节点之间的通信。尽管有一些研究揭示了简单数据分发协议的复杂动态，但数据分发过程的基本特性仍然难以捉摸。例如，很少有人知道物理协作中继策略、功率控制、MAC层调度和路由如何影响源的可实现传输速率、信息扩散的速度、数据传递的可靠性以及在地理区域上广播消息的能量。因此，数据传播协议的设计主要是以临时方式完成的。本项目主要研究数据分发协议的理论和系统设计，重点研究数据分发协议在无线传感器网络、无线Mesh网络等业务系统中的应用。该研究将开发一个统一的分析框架来表征不同物理层模型下随机和规则网络中数据分发过程的渐近性质和能量性能权衡。理论结果为跨层设计高效、健壮的数据分发算法奠定了基础。研究人员和学生将开发用于传感器网络管理(SNM)的一对多和多对一传输服务和API，以促进传感器网络中被管理设备和状态的配置和监控。

Data dissemination, i.e., transporting data or control messages from one to multiple end points, is an indispensable component in wireless network operations. Existing and emerging applications in multihop wireless networks, such as distributed industrial control, environmental monitoring, emergency response systems act on aggregates of wireless devices rather than communication between individual nodes. Although there exist a few studies that reveal the complex dynamics of simple data dissemination protocols, the fundamental properties of data dissemination process remain elusive. Little is known, for instance, how physical cooperative relay strategies, power control, MAC layer scheduling and routing affect the achievable transmission rate of the source, the speed of information diffusion, reliability of data delivery and energy to broadcast messages over a geographic area. Consequently, the design of data dissemination protocols is mostly done in an ad hoc fashion. This project focuses on the theory and system design of data dissemination protocols with emphasis on their applications in operational systems such as wireless sensor networks and wireless mesh networks. This research will develop a unified analytical framework to characterize the asymptotic properties and energy-performance trade-offs of data dissemination processes in random and regular networks under different physical layer models. The theoretical results serve as a foundation to the cross-layer design of efficient and robust algorithms for data dissemination. The investigator and students will develop one-to-many and many-to-one transport services and APIs for sensor network management (SNM) that facilitate the configuration and monitor of managed devices and states in sensor networks.