CAREER: Spatial Models and Algorithms for Sensor Networks

提案：0347400 INSTITUTION：得克萨斯大学奥斯汀分校PRINCIPAL调查员：沙克泰，桑杰标题：职业：传感器网络的空间模型和算法从购买食品杂货这样的日常任务，到通过河床追踪化学污染物，传感器网络技术将改变我们理解物理世界和与之互动的方式。这样的网络将支持通过数千个传感器节点的通信--在今天的网络中没有见过的规模的通信。受气体物理学的推动，为了理解压力或温度等宏观行为，单个原子的行为是无关紧要的，这个职业项目开发了一个基于传感器节点网络的连续统观点的研究计划--其中单个节点的属性聚集在一起，并仅表现为底层连续介质中的随机波动。该方案通过引入统计物理技术，建立了传感器网络算法设计和体系结构的框架。它被用来获得基本的见解，并为传感器网络通信开发算法-从查询、泛洪和路由设置等网络协议原语，到数据流路径优化，再到无线媒体上多个数据流的交互，以及它们对拓扑特性的影响。研究成果影响了广泛的传感器网络应用的设计方法，并被纳入德克萨斯大学奥斯汀分校的本科和研究生课程。通过得克萨斯大学奥斯汀分校的无线网络和通信集团工业附属公司计划，业界可以很容易地获得结果。一个集成了实际网络硬件和分析网络模型的网络试验台已经原型，用于测量和验证传感器网络算法，并进一步扩展到高中生。

PROPOSAL: 0347400INSTITUTION: U of Texas AustinPRINCIPAL INVESTIGATOR: Shakkottai, Sanjay TITLE: CAREER: Spatial Models and Algorithms for Sensor NetworksFrom everyday tasks like buying groceries, to tracking chemicalpollutants through a riverbed, sensor network technology willtransform the way we understand and interact with the physicalworld. Such networks will support communications over thousands ofsensor nodes -- communications of a scale not seen in today'snetworks.Motivated by the physics of gases, where the behavior of individualatoms are inconsequential in order to understand macroscopic behaviorsuch as pressure or temperature, this CAREER project develops a researchprogram based on a continuum viewpoint of a network of sensor nodes --where individual node properties aggregate, and manifest only asrandom fluctuations in an underlying continuous medium. By introducingtechniques from statistical physics, this proposal develops aframework for sensor network algorithm design and architecture. Thisis applied to obtain basic insights, and develop algorithms for sensornetwork communications -- from network protocol primitives such asquerying, flooding and route setup, to data flow path optimization, tothe interactions of multiple data flows over wireless media, and theirimpact on topological properties.Research results impact design methodology for a wide range of sensornetwork applications, and are incorporated into the undergraduate andgraduate curricula at UT Austin. The results are easily accessible toindustry through the Wireless Networking and Communications Groupindustrial affiliates program at UT Austin. A network testbed thatintegrates actual network hardware with analytical network models hasbeen prototyped to measure and validate sensor network algorithms, andfurthers outreach efforts to high school students.