面向凹型区域低密度部署的非测距节点定位算法研究

Range-free node localization algorithm plays an important role in the wireless sensor network node positioning. The resolved question will contribute to promote the development of wireless sensor network application service based on position. However, in the concave area and low density deployment environment, the current algorithms exist the metric of distance relation between neighbor nodes is not accurate enough. The distance error and localization error are also big. In view of the above problems, the project carried out the following research. First, the relation between the distance of neighbor nodes and intersection area of their communication range is researched. The calculation method of metric, which can accurately express the distance relation between neighbor nodes through Monte Carlo method and node compensation in the low density deployment environment. Then, analyzing the influence of the concave boundary on the shortest path between non-neighbor anchor nodes, a distance method, which can accurately estimate the distance between neighbor nodes, is designed by combining the metric and the method to filter out information between anchor nodes which is affected by concave boundary. Finally, node localization is considered as a multi-objective optimization problem. Utilizing the distance between neighbor nodes and connectedness between nodes, multi-objective functions are designed. The locations of unknown nodes are calculated by the Intelligent optimization algorithm. Relevant research progress and breakthrough of this project will provide theoretical evidence and algorithm support for range-free node localization algorithm to obtain high precise localization results in a concave area and low density deployment.

非测距节点定位算法在无线传感器网络节点定位中占有重要地位，这个问题的解决将有助于推动基于位置无线传感器网络应用服务的发展。现有算法在凹型区域低密度部署环境下存在邻居节点远近关系度量值不够准确，距离误差和定位误差较大的问题。针对上述问题，本项目将展开如下研究。首先，研究邻居节点距离和通信覆盖相交区域面积之间的关系，通过蒙特卡洛统计方法和节点补偿设计能够精确表达低密度节点部署环境下邻居节点距离远近关系度量值的计算方法；然后，分析凹型边界对非邻居锚节点之间最短路径的影响，结合度量值和过滤掉受凹型边界影响的锚节点信息，设计能够精确估计邻居节点距离的算法；最后，将节点定位看成多目标优化问题，用邻居节点之间估计距离和节点之间连通约束设计多目标函数，用智能优化算法完成未知节点的定位计算。本项目相关研究进展和突破将为非测距节点定位算法在凹型区域低密度节点部署中获得高精度节点定位结果提供理论依据和算法支撑。

对于无线传感器网络，在不知道自身位置的情况下，传感器节点发回的监测数据往往没有任何意义。受诸多环境因素的影响，节点部署区域会形成凹型区域，而且节点的部署具有低密度的特点，因此研究如何在凹型区域低密度部署环境下，用非测距节点定位算法获得未知节点的精确定位结果具有重要的研究意义。获得的主要研究成果如下：1）用邻居结点之间的几何关系，非邻居结点之间的跳数和锚节点的坐标，同时考虑凹型边界的影响，提出了一种面向凹型区域的无线传感器网络高精度邻居节点距离估计方法。2）为了获得更好的定位结果，本项目将未知节点位置计算问题看成最优化问题，对智能优化算法进行了研究。针对人工免疫系统在处理动态优化问题时会面临特殊的挑战的问题，提出了一种基于固有和适应性反应及记忆机制的人工免疫算法。针对传统克隆选择算法在求解复杂最优化问题时会遇到早熟收敛和易陷入局部最优的缺点，提出了基于成功历史自适应的混合克隆选择算法。通过对免疫克隆机制进行深入的研究，考虑多重信息融合的思想，提出了一种新的基于基因重组和改进超编译算子的克隆选择算法。项目组还在无线传感器网络身份认证、无线视觉传感器网络、信号处理、大数据分析等方面取得了一定研究成果。项目组共发表论文12篇，其中SCI收录3篇，EI期刊6篇，申请国家专利3项。上述研究成果对基于位置的无线传感器网络应用提供了理论和技术支撑。