【作者】 王莹；

【导师】 George Almpanidis； 凡高娟；

【作者基本信息】 河南大学 ， 计算机应用技术， 2019， 硕士

【摘要】 无线传感器网络(Wireless Sensor Networks,WSN)部署在人迹罕至和环境恶劣的区域,经常会面临节点失效和网络抗毁性差的问题,既缩短了网络生命期,又严重影响了网络正常的使用,致使既定的网络监测任务无法按时完成。因此,网络的能耗和抗毁性能成为了WSN拓扑设计时需要考虑的关键问题。无标度网络中节点度分布不均匀且具有幂律特性,故节点对随机打击具有很好的抗毁性,可以满足WSN拓扑设计的要求。目前大多数无标度网络模型都是基于BA(Barabási-Albert)模型进行扩展演化,但是BA模型择优连接的单一性和增长模式与实际网络存在较大差异,具有一定的局限性。因此,本文在现有无标度模型的基础上,再结合WSN的实际特征,设计出具有能耗优化和抗毁性的网络模型。本文的主要研究内容如下:1.针对WSN中常出现的节点失效的问题,提出了基于Possion增长的加权无标度拓扑演化模型(Poisson growth Weighted Scale-free topology evolution Model,PWSM)。该模型基于BBV(Barrat-Barthelemy-Vespignani)加权网络模型将动态权值引入到网络的拓扑演化过程中,同时在优先连接机制中不仅考虑了经典无标度BA模型中节点度的因素,而且把节点的局域性、节点的剩余能量综合考虑,并定义了新的边权表达式。通过理论推导,证明由此模型演化生成的拓扑点权、边权和节点度均服从幂律分布,具有无标度特性。仿真结果表明该网络拓扑能够均衡网络能耗,具有较好的容错性。2.针对WSN在实际应用中面对蓄意攻击时的脆弱性等问题,提出了具有可控幂律的无标度拓扑演化模型(Controllable power Rate Scale-free topology evolution Model,CRSM)。CRSM模型首先通过HEED(A Hybrid,Energy-Efficient Distributed clustering approach)分簇算法实现簇头的分布,其次通过初始网络中簇头节点随机行走的方式构建了局域世界,最后考虑节点度,节点剩余能量和节点间距离,使剩余能量大且距离短的簇头节点具有较高的连接概率。通过理论推导,证明该模型生成的网络拓扑度分布服从幂律分布,且由于择优连接概率中调节参数的引入,使得幂律指数在一定范围内可调。3.为了使CRSM模型演化后的拓扑具有很好的自我修复、可重构性和可调性,在CRSM的基础上提出了具有链路补偿机制的无标度拓扑演化模型(Link Compensation mechanism in Scale-free topology evolution Model,LCSM)。该模型根据WSN变化的实际特征加入了链路删除和链路补偿机制,通过链路删除机制删除网络中能量小,距离大的节点,且相应的链路也被删除,可以降低网络能耗;通过链路补偿机制对于失效链路进行了适当补偿,延长了网络生命期。仿真结果表明,CRSM和LCSM模型演化的网络拓扑在随机和蓄意打击下都具有很好的抗毁性和容错性,可满足恶劣环境对WSN鲁棒性的要求。更多还原

【Abstract】 Wireless sensor networks(WSN)are deployed in areas with inaccessible and harsh environments.They often face problems of node failure and poor network intrusion,which greatly shortens the network life cycle and seriously affects the normal use of the network,tasks cannot be completed on time.Therefore,the energy consumption and damage resistance of the network is a key issue to be considered when designing the WSN topology.The node degree distribution in a scale-free network is not uniform but has power-law characteristics.Therefore,nodes have good invulnerability to random strike and can meet the requirements of WSN topology design.At present,most scale-free network models are based on the Barabási-Albert(BA)model for extended evolution,but the unity and growth model of the BA model’s preferred connection is quite different from the actual network,which has certain limitations.Therefore,based on the existing scale-free model,combined with the dynamic requirements of WSN,this thesis designs a network model with energy optimization and invulnerability.The main research contents of this thesis are as follows:1.Aiming at the problem of node failure that often occurs in WSN,a weighted scale-free topology evolution model Poisson growth weighted scale-free topology evolution model(PWSM)based on Poisson growth is proposed.This model introduces the dynamic weights into the topology evolution of the network based on the weighted network model proposed by Alain Barrat et al.(BBV for short).At the same time,in the priority connection mechanism,not only the factor of the node degree in the classical scale-free BA model is considered,but also the locality of the node and the residual energy of the node.Considered comprehensively,a new edge weight expression is defined.Through theoretical deduction,it is proved that the topological point weight,edge weight and nodal degree generated by the evolution of this model obey the power law distribution and have scale-free characteristics.The simulation results show that the network topology can balance network energy consumption and has high transmission efficiency and connectivity.2.Aiming at the vulnerability of WSN to deliberate attacks in practical applications,a scale-free topology evolution model with controllable power law controllable power rate scale-free topology evolution model(CRSM)is proposed.This CRSM model firstly realizes the distribution of cluster heads with the help of the Hybrid Energy Efficient Distribute(HEED)clustering algorithm.Then,when constructing the inter-cluster topology,the node degree,the residual energy of the cluster head node and the distance between nodes are considered,so that the cluster head nodes with large residual energy have higher connection probability.Finally,the topology evolution between cluster heads is carried out by combining scale-free networks with random walks.The theoretical derivation shows that the network topological degree distribution generated by the model obeys the power law distribution,and the power law index is adjustable within a certain range due to the introduction of the adjustment parameters.3.In order to make the evolved topology have good self-healing,reconfigurability and adjustability,a scale-free topology evolution model with link compensation mechanism link compensation mechanism in scale-free topology evolution model(LCSM)is proposed based on CRSM.This model adds link deletion according to the actual characteristics of the WSN change.For nodes with low energy,large distance nodes and their corresponding links are also deleted,which can reduce network energy consumption and the cost of joining links for failed links and nodes.Thus,the compensation mechanism can extend the life of the network.The simulation results show that the network topology evolved by the above two models has high robustness under random and deliberate attacks,which can meet the robustness requirements of WSN in harsh environments.更多还原