【作者】 王颖；

【导师】 江国星；

【作者基本信息】 华中科技大学 ， 通信与信息系统， 2015， 硕士

【摘要】 无线传感器网络(Wireless Sensor Networks,WSNs)是一个由功率受限的传感器所组成的以无线多跳的方式进行感知和通信的自组织网络。无线传感器网络在军事行动、环境控制、智能家居、医疗保障及其它商业方面具有广泛应用,从而备受不少研究者的瞩目。在执行关键任务的时候,维持无线传感器网络的连通性至关重要,因为各个传感器节点必须相互通信才能对所检测到的事件迅速做出最理想的反应。然而,传感器节点的失效可能会使它与其它传感器节点失去联系甚至造成网络分离现象,这样会导致整个网络无法正常运行。无线传感器网络通过修复连通性维持着网络连通的状态,从而保障了网络的正常运行。本文在分析无线传感器网络中主流的连通性修复算法的基础上,提出一个基于Fermat点的分布式连通性修复算法CRFP(Connectivity Restoration based on Fermat Points)来解决来传感器节点失效导致的连通性问题。CRFP算法预先基于连通支配集CDS(Connected Dominating Set)判断失效节点是否是关键节点,并且通过把失效节点的邻居节点移动至最合适的Fermat点位置的方式来修复连通性。CRFP算法追求的是传感器节点的优化重组,并且实现以较小的开销修复连通性的目的。由于CRFP算法只能处理单节点失效的情形,因此我们在此基础上提出在多节点失效情形下的连通性修复算法MCRFP(Multiple CRFP),用于处理在出现多个关键节点失效情况的连通性修复问题。本文分别从理论和仿真两个方面对CRFP算法和MCRFP算法的性能进行分析,并且在网络仿真器NS2(Network Simulator version 2)上实现了算法的仿真。通过分析实验的数据,验证了CRFP算法和MCRFP算法的性能。更多还原

【Abstract】 Wireless Sensor Networks(WSNs) is composed of a group of power-constrained sensor nodes with functions of sensing and communication via wireless multi-hop method. Due to the great application on military, environment, health, household and other commercial applications, WSNs have successfully got considerable attention from scientific researchers. Maintaining the connectivity in WSNs is very important in mission-critical applications where sensors have to quickly plan optimal response to detected events. However, the failure of one sensor may lead to the disconnection with others and even the network partitioning, thus making the network dysfunctional. The connectivity restoration is particularly important for the network in order to recover back to the original state.This thesis analyzes the current main connectivity restoration algorithms for WSNs and presents a distributed algorithm CRFP——Connectivity Restoration based on Fermat Point to tolerate the connectivity issue caused by the failure of sensor node. CRFP proactively identifies the sensor is critical node or not to the network connectivity based on Connected Dominating Set(CDS) in advance, and restores the connectivity by replacing the locations of Fermat points with the suitable neighbors of the failed node. It pursues an optimized rearrangement of sensor nodes and achieves the goal of recovering the connectivity with a minimized overhead. CRFP deals with the circumstance where only one node fails. Therefore, we improve it and propose MCRFP(Multiple CRFP) to deal with the circumstance where more than one node fail.The thesis analyzes the performance of CRFP and MCRFP theoretically and experimentally. The thesis realizes the simulation on NS2(Network Simulator version 2), and verifies the performance of CRFP and MCRFP.更多还原