分布式传感网高效信息融合安全一致性研究

Information fusion of distributed wireless sensor networks has wide applications in the civil and military fields. However, the lack of a centralized entity that monitors the activity of all the sensor nodes makes wireless sensor networks be prone to external malicious attacks. Therefore, it is of importance to study the security of information fusion of wireless sensor networks. For consensus-based information fusion of wireless sensor networks with communication delays, measurement noises and Byzantine attacks, in this proposal we plan to carry out the following research. First, we will propose the consensus-based information fusion algorithms of wireless sensor networks with communication delays, measurement noises and Byzantine attacks, analyze whether the algorithms achieve consensus and validity, and find the conditions guaranteeing the achievement of consensus and validity. Second, we will design novel robust consensus-based information fusion algorithms to reduce the dependence of fault tolerance to Byzantine attacks on the network connectivity. Third, we will develop finite-time robust fault detectors and identifiers, the goals of which are to fast and accurately detect and identify the sensor nodes subject to Byzantine attacks and then to carry out fault isolation by changing the network topology actively. Fourth, we will investigate the security of consensus-based information fusion of wireless sensor networks with dynamic topologies. This project is the application-oriented theory research, and might provide the safety basis for the application of information fusion of wireless sensor networks.

分布式无线传感网信息融合在民用和军用领域都有着广泛的应用，但由于缺乏监控所有传感器节点活动的中心实体，网络容易遭受外部的恶意攻击，因此研究网络信息融合的安全性问题有着重要的意义。本项目针对拜占庭攻击下含通信时滞和测量噪声的无线传感器网络一致信息融合，拟开展以下研究内容：（1）提出融通信时滞、测量噪声及并发传感器节点拜占庭攻击的无线传感器网络一致性信息融合算法，分析其是否实现一致性与合法性，寻找实现一致性与合法性的条件；（2）设计新的鲁棒一致性信息融合算法，降低抗拜占庭攻击的容错能力对网络拓扑结构连通性的要求；（3）设计有限时间鲁棒故障检与辨识器，快速准确检测与辨识拜占庭攻击传感器节点，进而主动改变网络拓扑结构以实施故障隔离；（4）研究动态网络拓扑结构下的无线传感器网络一致信息融合的安全性。本项目属于应用基础研究，为分布式无线传感网信息融合在实际中的应用提供安全性基础

随着工业4.0时代的到来和信息物理社会的日趋成形，现代的工业工程和先进制造中所面对的无线传感器网络系统通常网络特征信息复杂、规模庞大，这对无线传感器网络系统的分析和算法设计带来了极大的困难。另一方面，网络数据通讯中的传输延迟、数据包丢失、通讯故障、量化误差等问题给无线传感器网络系统的分析提出了新的困难和挑战。.本项目研究了随机测量数据丢失及切换拓扑下的分布式传感器网络模型，基于传感器节点自身信息及邻居节点信息，提出分布式传感器网络一致性信息融合算法，实现了对目标的状态估计；研究了传感器节点受到网络随机攻击和传输信道固有因素的影响，引入弹性策略，基于均方稳定性判据，设计滤波器保证滤波误差系统稳定；研究了一种量化的分布式无模型自适应迭代学习双向一致性控制算法，降低通信时滞、测量噪声、传感器故障对于系统的影响；研究了一种事件触发分布式数据驱动多智能体系统双向编队控制算法，基于观测者的事件触发机制，提高了网络通信资源的利用率。.本项目发表(含录用)国际SCI期刊论文19篇(第一作者/通讯作者论文16篇)；EI收录3篇；申请发明专利6项，获得授权5项(美国发明专利1项)；获得软件著作权8项。已发表的研究成果得到海内外学者的引用及推广。此外，项目进行过程中，教师和学生通过选题、开题、中期、毕业答辩以及论文、专利、软件著作权等写作训练和系统开发中获得了宝贵的科研协同经验和深厚学术素养，为今后更大规模科研工作打下基础。