【作者】 胡一帆；

【导师】 丁永生；

【作者基本信息】 东华大学 ， 控制理论与控制工程， 2012， 博士

【摘要】 无线传感器网络是由部署在监测区域内的大量微型、低成本传感器节点组成的自组织多跳网络系统。这种网络的易部署性、自适应性特点,使得它在环境监测、精细农业、远程医疗等领域有着广泛的应用前景。但由于数目庞大的节点硬件结构脆弱且能量受限、部署环境恶劣等特性,网络经常出现节点损坏、通信链路断开等故障。近年来传统基于冗余的容错路由技术由于没有综合考虑能量受限等无线传感器网络特性,如何设计节能高效的容错技术、保证数据传输的鲁棒性以提高网络性能已经成为无线传感器网络研究的关键问题。本论文针对无线传感器网络中的路由容错问题,借鉴生物协同进化智能算法的优势,同时考虑到路由优化过程中能量平衡机制,通过研究无线传感器网络中分簇、异构节点、移动的汇聚节点(Sink)等网络结构的容错路由问题,建立了相应的路由容错模型,解决在无线传感器网络中的路由容错优化等复杂计算应用问题。取得的研究成果包括如下几个方面：(1)以分簇的无线传感器网络为场景,针对异构网络簇内路由失效后的路径恢复问题,提出了基于路径编码的无线传感器网络簇内的路由容错模型及节点失效模型。综合运用免疫协同进化粒子群算法和簇内多路径路由策略,构建异构网络的簇内路由智能容错机制,使得路由容错模型的路径搜索能集中在高质量的解搜索空间内。仿真实验证明了该策略提高了簇内数据传输的效率和容错能力。(2)对异构无线传感器网络的簇间超级节点之间的容错路由问题进行了进一步分析,建立了相应的簇间簇内路由智能容错模型,运用主从粒子群协同进化免疫算法研究最优的替代路由构建策略,通过路径编码、主从群协同更新进化、克隆复制、高频变异、克隆选择等操作进行问题求解,以提高算法的运行效率和反应能力,从而提升异构无线传感器网络的整体容错性和数据传输的可靠性。最后通过实验对该算法及理论分析进行了验证。(3)针对现有的单移动Sink无线传感器网络协议复杂、通信开销大而不能高效适应拓扑频繁变化的问题,建立了单移动Sink路由恢复模型,并采用正交免疫粒子群算法来维护随着Sink移动而变化的拓扑路由,同时有效地降低通信开销,减少网络能耗。实验表明采用该策略的单移动Sink网络能量分布更均衡,容错性能也得到了相应的提升；(4)对多个移动Sink无线传感器网络的路由容错进行了进一步分析,结合多个Sink移动的特点和网络节点失效问题综合设计网络路由智能容错模型,并采用内分泌粒子群协同进化思想来构建高效可靠的替代路径,通过路径编码、激素群选择、粒子群协同更新等操作来求解问题。仿真结果也进一步验证了该策略能提高数据流传输的鲁棒性,均衡化网络能量消耗,延长网络生存周期。最后,总结了论文的研究内容,指出了研究中存在的不足,展望了下一步的研究方向。更多还原

【Abstract】 Wireless sensor networks (WSNs) consist of a large amount of inexpensive, miniature sensor nodes which are deployed in monitoring region, forming a self-organized multi-hop network system. The features of easy deployment, self-adaptive and low cost enable WSNs to conduct many applications, such as precision agriculture, environment surveillance and remote medicine. However, there are often some faults happened, such as node failure, wireless link breakage, which is because the massive nodes in WSNs are energy constrained and usually deployed in harsh environments. As the feature of energy restriction and others have been rarely considered in recent years, the tranditional routing fault-tolerant protocols based on redundancy are difficult to efficiently achieve in WSNs. Designing efficient fault tolerant technology which can ensure the robustness of data transmission, prolong the network lifetime and improve its performance has been a critical issue on WSNs research.In this study, the fault-tolerant routing problem of WSNs is focused, based on the superior features of biological co-evolution intelligent algorithm, and considering the energy balance mechanisms in process of routing optimization. The fault-tolerant mechanism of heterogeneous nodes and mobile sink nodes is studied by building routing models, and applying in the route optimization of complex computation applications in WSNs. This paper focuses on analyzing the fault-tolerant routing technology of clustering and mobile sink structure, and the main contributions of the paper are as follows:(1) Based on heterogeneous WSNs and path coding, the intra-cluster fault-tolerant routing model and node fault model are proposed for the path recovery problem after intra-cluster routing failure. With immune co-evolution particle swarm algorithm and multi-path routing method, we build the intra-cluster routing intelligent fault-tolerant mechanism to make the path searching concentrating on high-quality solution within the search space and improving the algorithm efficiency and responsiveness. The simulation results have shown its efficiency and fault-tolerant cability of intra-cluster data transmission comparing with the classical clustering protocols(2) The inter-cluster fault-tolerant routing problem of heterogeneous WSNs is further analyzed. We build the inter-cluster route, use the master-slave co-evolution immune particle swarm algorithm to investigate the optimal alternative routing strategies, and solve the problem with path coding, master-slave swarm updating, cloning, high-frequency mutation, clonal selection and other operations, which can improve the overall fault tolerant ability and reliability of inter-cluster and intra-cluster data transmission. Finally, theoretical analysis of the algorithm was verified through experiments.(3) Some existing routing protocols of mobile sink WSNs are not suitable for the problem of topology frequent change, due to their complex computing and high communication cost. We adopt a single mobile sink WSNs routing recovery model, and use the immune orthogonal learning particle swarm algorithm to repair the routing topology changed by the sink movement. The method can also reduce the communication overhead and network energy consumption of single mobile sink WSNs. Experiments show that SMS-WSNs have a more balanced energy distribution, and the improved fault tolerant performance using this strategy.(4) We further analyze the fault-tolerant routing of multiple mobile sink WSNs and design the routing model, considering both the characteristic of multiple sink mobility and network node failure problem. Then we use the endocrine co-evolution particle swarm algorithm to build the efficient and reliable alternative path, and solve the problem with path coding, hormone group selection, swarm co-evolution updating and other operations. The simulation results have also validated that the strategy can improve the robustness of data stream transmission, distribute the network energy consumption, and prolong the network lifetime.At the end, we summarize the content, advantage and deficiency of the paper, and narrate further development of the study.更多还原