面向入侵目标监测的无线传感网自愈节能技术研究

The main task of wireless sensor network (WSN) is to accurately acquire valuable physical world information, expand the perception range of humans, help people understand and sense their surroundings in real time. Energy constraint is an important issue for widely applications. Monitoring time-varying and diversity characteristic of environment lead to nodes failures, which cannot guarantee monitoring accuracy, especially for unpredictable environment application, and worsen network monitoring performance. Intrusion target monitoring application is used as a background in this project research, and a node scheduling self-healing technology is research priorities, which includes four parts. Firstly, character parameters of the intrusion target will be analyzed, and then a measurement of network lifetime will be presented to ensure event monitoring in real time and agility. Secondly, coverage hole diffusion problem model is establishment based on dynamic Bayesian network theory, which can provide guiding principles for self-healing technology in node scheduling schemes. Thirdly, a self-healing strategy is designed in node scheduling to improve network robustness according to the circumstance that node location information is free and location information is demand. Finally, a Gini index is used in economics as a coverage evenness measurement for intrusion target monitoring application, comprehensive index merit of performance evaluation is designed, which is used to evaluate performance of self-healing strategy in node scheduling. Validity and feasibility of the proposed theory is measured by intrusion monitoring instance. Research results have important theoretical significance and reference value for development and popularization of wireless sensor network in many new application fields.

无线传感器网络的主要任务是准确获取客观物理世界有价值的信息，拓展了人类的感知范围,帮助人们了解并实时感知周围环境。能量受限是传感器网络广泛应用面临的首要问题，监测环境的时变性和多样性导致节点失效，难以保证监测精确性，尤其应用环境的不可预知性，更加恶化网络监测性能。本项目以入侵目标监测应用为背景，研究节点调度自愈技术：分析入侵目标特征参数，建立网络生存度量模型，保证事件监测的敏捷性和实时性；利用动态贝叶斯网络理论，建立节点调度覆盖洞扩散问题模型，为节能自愈技术提供指导原则；设计节点位置信息已知和未知情况下的节点调度自愈策略，提高网络健壮性；利用经济学中的基尼系数作为目标监测的均衡度，设计基于模糊理论的网络综合性能评价指标，分析评估节点调度自愈策略性能。用目标监测实例验证所提理论的正确性与可行性。研究结果对促进无线传感器网络在更多新型应用领域的发展和普及具有重要的理论意义和参考价值。

本课题针对能量受限的无线传感器网络节点调度机制中存在的监测精确性难以保证问题，以入侵目标监测应用为背景，展开无线传感器网络自愈节能技术。（1）分析节点的自身节能特性，给出节点在调度中存在的不转发数据、过度休眠和抢占休眠三种行为，实验分析节点自私行为在覆盖和能量方面对邻居节点的影响，并从网络覆盖率、网络生存时间、网络能量消耗均衡性等方面分析节点的自私行为对整个网络性能的影响；（2）利用复杂网络无标度演化模型，分析网络的健壮性与抗毁性，提出DEAM(Distance Energy Average-degree Model)演化模型，从节点度、剩余能量、传输距离、网络平均度等因素对网络的健壮性与抗毁性进行分析；（3）提出地理位置信息未知的覆盖漏洞发现算法CHDA (Coverage Holes Discovery Algorithm)，以节点为极点建立的极坐标中计算出相邻节点的相对位置信息和该节点被其邻居节点覆盖的边缘弧的信息，并且根据节点的单纯覆盖弧序列的定义计算出节点被其邻居节点覆盖的单纯覆盖弧序列，从而得到网络中的覆盖漏洞，并在此基础上进行以入侵距离为修补约束的覆盖洞修补策略设计；（4）设计节点位置信息已知的节点调度自愈方法，根据节点剩余能量和监测覆盖要求对节点状态进行调度，尽可能避免或延缓覆盖空洞问题的产生，并且设计休眠时间动态调整调度方案，从而达到延长网络生存时间的目的；（5）利用基尼系数（Gini index）、覆盖率、能量均方差、入侵时间、入侵距离等评估参数，设计EvenCover++算法，提高网络能量消耗均衡性和覆盖均匀性。研究成果对于促进无线传感器网络在更多新型应用领域的发展和普及具重要的理论意义和参考价值。