面向渐变事件的无线传感器网络监测及评价验证方法研究

项目名称： 面向渐变事件的无线传感器网络监测及评价验证方法研究

项目编号： No.61272537

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 杨武

作者单位： 哈尔滨工程大学

项目金额： 80万元

中文摘要： 本课题研究基于无线传感器网络的面向生物化学气体、药剂泄露等渐变事件的监测技术及其相关实验验证方法。研究内容包括：适用于事件监测的传感器网络结构建立与维护、渐变事件容错检测、渐变事件边界识别与跟踪以及事件监测性能分析与仿真等。本课题的研究以实时、节能、可靠的检测和跟踪渐变事件为目标，首先引入"中介点"的概念，建立适用于事件监测的簇型网络结构以及簇间多路径路由；在事件检测阶段，根据簇内节点监测数据的时空相关性，采取层次式多元统计分析方法建立系统化的错误样本识别框架，并由簇头结点利用轻量级分类器对监测数据进行融合决策；在渐变事件跟踪阶段，利用高斯混合模型识别簇内局部事件边界，再由sink节点完成局部边界的连接，并采用基于主从节点的预测技术和相关睡眠唤醒机制实现对渐变事件边界的跟踪。本课题在无线传感网网内信息协作处理、事件检测、事件边界识别和跟踪技术上具有重要的理论意义和实际应用价值。

中文关键词： 无线传感器网络；信息协作处理；事件监测；事件检测；边界跟踪

英文摘要： This subject studies the monitoring technology and its related experimental verification method of gradual events, such as biochemical gas and agent leakage in wireless sensor network. The study includes: network establishment and maintenance for event monitoring, fault-tolerant event detection, gradual events boundary identification and tracking, event monitoring performance analysis and simulation. We set real time, energy conservation, reliable detection and tracking of gradual events as our goal. First of all, we use "intermediary points" to establish cluster-based network structure and multi-path routing strategy of adjacent clusters for event monitoring. In the event detection stage, according to the spatial and temporal correlation of monitoring data of intra-cluster nodes, we use hierarchy multiple statistical analysis method to establish systematic error sample identification framework，and then the cluster head nodes use lightweight classifier to make fusion decision of monitoring data. In the gradual events tracking phase, we use Gaussian mixture model for cluster internal local events boundary identification, then use sink nodes to connect local boundaries, and use master-slave nodes forecasting technique and sleep-arousal mechanism to achieve the tracking of event boundaries. This subject has importa

英文关键词： wireless sensor network；collaborative information processing；event monitoring；event detecting；boundary tracking