面向海洋环境监测的移动传感网络密钥管理与安全拓扑控制研究

Marine mobile sensor networks have been received great attention in the field of marine environment monitoring due to its stronger flexibility of sensing. The mobility of the nodes leads to frequent network topology changing, the uncertainty of the communication link, and the possibilities that the nodes can be captured or the communication data can be intercepted and tampered. The dynamics of network topology has brought more questions for security application of sensor networks. It has more important practical significance and application value to explore the effective methods of key management and optimization of safety topology control with support for dynamic topology in marine mobile sensor networks. According to the principle of the spatial correlation of underwater object motion and the mobility prediction based on the cooperative mechanism, this project plans to make the estimation to the motion state of the unknown nodes, and get the completed dynamic topology information of the network. The tree-based path-key establishment can make sure the key connectivity, and the safety dynamic key management can be realized by the discrepant keys establishment and the improved Diffie-Hellman key agreement mechanism. The dynamic clustering algorithm based on multi-objective optimization and the introducing of particles competition mechanism is established by comprehensively considering the factors of nodes energy, distance, mobility and the capture probability. The problems of continuity of data communication while nodes moving across the clusters and nodes dynamic security identity authentication can be solved by the soft handoff mechanism. The performance of the proposed mechanism and algorithms can be evaluated and compared by the simulation experiment and theoretical analysis.

海洋移动传感网络由于其较强的感知灵活性，在海洋环境监测领域正获得极大关注。节点的移动性导致网络拓扑变化频繁、通信链路不稳定、节点存在被捕获或通信数据存在被拦截、篡改的可能。拓扑动态性给传感网络的安全应用带来许多问题，探索研究动态拓扑下海洋移动传感网络密钥管理及安全拓扑控制的有效方法具有重要现实意义和应用价值。本课题拟根据水下物体运动的空间相关性原理及基于协作机制的移动预测方法对未知节点进行运动状态估计，获取网络完整的动态拓扑信息；基于树的路径密钥建立方法来保证密钥连通性，并通过差异密钥建立方法及改进的Diffie-Hellman密钥协商机制实现安全动态密钥管理；综合考虑节点能量、距离、移动和捕获概率因素构造基于多目标优化、引入粒子竞争机制的动态分簇算法；通过软切换机制解决节点跨簇移动的数据通信连续性及节点动态安全身份认证问题。通过模拟实验和理论分析的手段评价和比较所设计机制与算法的性能。

海洋移动传感网络由于拓扑变化频繁、通信链路不稳定、数据通信安全易受威胁等原因，使得构建广覆盖、强安全的有效、稳定的动态监测网络面临技术挑战。本课题面向海洋移动传感网络，研究网络安全通信的密钥管理与安全拓扑控制关键技术：提出一种综合考虑节点能量因素和节点捕获概率的无线传感网络密钥管理方法，采用EBS（Exclusion Basis Systems）结构，有效实现簇间组密钥的更新；基于改进的LKH（Logical Key Hierarchy）和Iolus密钥管理方案，提出一种针对异构网络的组播密钥管理方案；进而以基于Merkle树思想改进的LKH密钥管理方案和SKDC（Simple Key Distribution Center）密钥管理方案为基础，提出一种可降低动态组播密钥管理的通信、计算和存储消耗的大规模海洋异构无线传感网络组播密钥管理方案，使之更加适应于海洋网络的高动态性。提出了一种在全局搜索能力和收敛速度上都有显著提升的网络拓扑优化算法：FW-PSO，无论节点初始负载调整参数α取何值，所提出算法优化后的网络不仅在度大和度小攻击策略下的关键阈值Tc减小的百分比最多，而且随机攻击下的网络连通性提高的百分比最多，网络的静、动态抗毁性获得显著增强；针对网络攻击检测，提出基于改进的遗传算法和深度置信网络相结合的传感网络攻击检测模型及基于节点行为分析和云模型的传感网络信任安全方案；针对网络动态拓扑恢复，提出一种基于有限状态机与网络连通支配集的分布式传感网络连通性修复算法；进而又提出一种混合式的面向网络感知覆盖率的连通性恢复算法，旨在用最小的覆盖面积损失开销恢复网络连通性。针对海洋传感网络的安全路由，提出一种基于强化学习的海洋传感网络水下机会路由协议，能够灵活适应海洋动态网络环境，具有数据传输可靠性高、鲁棒性强、时延低、能量利用率高的特点。研究成果均通过模拟实验获得了充分的验证分析。