无线传感器网络的数据丢失与重构研究

项目名称： 无线传感器网络的数据丢失与重构研究

项目编号： No.61303202

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 孔令和

作者单位： 上海交通大学

项目金额： 23万元

中文摘要： 无线传感器网络在传输过程中，数据包的丢失普遍存在，导致收集到的数据集不完整。然而很多基础科学研究比如自然环境探索、农业信息监控、野生动物观察，都依赖于完整准确的传感数据集来了解对象的全貌。如何在数据丢失不可避免的情况下，尽可能提高数据集的完整性具有重要的研究意义。因此本课题提出对无线传感器网络数据丢失和重构的研究。无线传感器网络的数据丢失有其特殊性质，比如：无线信道的不稳定性、多跳拓扑结构的干扰、数据爆发的拥塞、节点失效等原因。使得传统插值方法在无线传感器网络的数据重构中表现不佳。我们针对无线传感器网络的特性，分别从建立数据丢失数学模型；探索数据重构方法；设计数据收集协议；搭建实验和仿真平台等四个方面对本课题进行深入分析，目标是将数据重构的准确性/数据丢失比提高一个阶。

中文关键词： 无线传感器网络；数据丢失模型；数据收集协议；数据重构；

英文摘要： Data loss is ubiquitous in the transmission of wireless sensor networks, which results to an incomplete sensory data set. However,a lot of basic scientific works, e.g., nature discovery, agriculture information surveillance, and wildlife observation, heavily rely on the accurate and complete sensory data set to know the full view of the objects. It is significant to study how to increase the completeness of a data set in the case of inevitable data loss. Hence, this project propose to do the research on data loss and reconstruction in wireless sensor networks. Data loss in sensor networks has its special patterns due to instability from wireless channels, collision from multi-hop tree/cluster topologies, congestion from burst data, and unexpected nodes' failure. Thus, the conventional interpolation methods perfom poorly in data reconstruction of sensor networks. Based on the feature of sensory data loss, we plan to deeply study this project from four dimensions: to mine the data loss patterns, to discover the data reconstruction methods, to design the data collection protocols, and to build the experiment and simulation platforms. Our goal is to increase at least one order of the accuracy/loss ratio of data reconstruction.

英文关键词： Wireless Sensor Networks；Data Loss Pattern；Data Collection Protocol；Data Reconstruction；