大规模无线传感器网络定位误差分析及优化方法研究

Due to cost constraints, a large-scale wireless sensor network (WSN) normally consists of a mass of cheap sensor nodes and a small portion of anchor nodes with known locations, and obtain inter-node distance estimates based on RSS (Received Signal Strength) measurements, which significantly restricts the performance of localization algorithms. Hence, localization in large-scale WSNs has become an attractive research topic in the field of WSNs. This project will deal with the mechanism of sensor localization in large-scale WSNs, establish the scaling laws of and the propagation rule of multihop sensor localization errors, and then reveal intrinsic mechanism of as well as the propagation rule of multihop localization errors. And on this basis, by investigating the influence of anchor placement on localization accuracy, the dependency between the optimal anchor placement and the density of sensor nodes is demonstrated, and further, the optimal algorithm for placing anchors is presented. After carefully studying the key parts of distributed localization algorithms, including estimating inter-node distances based on RSS, localizing a single sensor node and flooding messages from anchor nodes, a systematic optimization framework is taken into consideration based on information fusion and graph theory, with the result that an energy-efficient distributed localization algorithm is proposed to restrict the propagation of localization errors. As a result, the outcomes of this project are helpful to the error theory of multihop localization in large-scale WSNs and provide theoretical guidance and practical methods for developing and deploying large-scale WSNs.

受成本因素的限制，大规模无线传感器网络通常由大量廉价节点和少量位置已知的锚点组成，并采用基于RSS（Received Signal Strength）测距方法，严重制约了定位算法的性能，已成为该研究领域亟待解决的关键问题。本项目将通过分析大规模无线传感器网络的定位机制，建立多跳定位误差的标度率及误差传播公式，从而揭示多跳定位误差的形成机理和传播规律；探索锚点位置对定位精度的影响规律，揭示锚点最优部署位置与节点密度的依赖关系，进而研究基于随机网络模型的锚点部署优化算法；深入分析分布式定位涉及的RSS测距、单个节点定位和洪泛广播三个关键环节，采用信息融合和图论等方法研究面向大规模无线传感器网络的多层次定位性能优化方法，形成具有节能和误差传播抑制功能的分布式定位算法。本项目的研究成果有利于完善多跳定位的误差理论，并为开发和部署大规模无线传感器网络提供理论指导和实用方法。

大规模无线传感器网络通常由许多能量受限的廉价节点和少量位置已知的锚点组成，并采用基于RSS（Received Signal Strength）的节点间测距方法实现节点定位。本项目旨在研究并解决大规模无线传感器网络中无专用设备的节点定位方法所面临的精度有限、通信代价高的难题。..首先，研究了大规模无线传感器网络中节点定位误差的传播规律。通过引入随机网络模型，建立了节点多跳定位的误差模型，揭示了定位结果的均方根误差随着距离锚点的最小跳数增大而线性增加，并随着网络中节点密度的增大而减小。基于CRLB（Cramer-Rao Lower Bound）和三边定位算法的大量仿真实验验证了以上结论。该研究从理论上揭示了节点定位误差传播的基本规律，可用于评估节点位置估计的可靠性，有助于开发大规模无线传感器网络应用。..其次，利用基于距离相关噪音模型的单跳节点定位场景，探究大规模传感器网络中多跳定位场景中锚点的最优部署策略。利用FIM（Fisher Information Matrix）行列式作为评价度量标准，建立了锚点最优部署的充分必要条件，以及在随机网络模型中锚点最优部署的必要条件。该研究不仅具有坚实的理论基础，而且具有较强的实用性，有助于指导在大规模无线传感器网络中部署锚点。..再次，根据节点间的RSS和邻接信息，采用MLE（Maximum Likelihood Estimator）实现了一种RSS测距校正方法，并给出了相应的CRLB，并通过仿真和真实数据验证提出方法的有效性。该方法提高了无专用设备的节点测距精度，有助于推动各种依赖于节点间距离的传感器网络应用。..最后，提出了基于FIM的锚点洪范控制策略，结合前述的RSS测距校正算法，实现了一个改进的大规模传感器网络节点定位算法。仿真实验表明该方法在定位精度和通信能耗两个方面均优于已有算法。该方法对于设计和部署大规模无线传感器网络具有重要的指导意义。..综上所述，本项目的研究成果有利于完善大规模无线传感器网络中多跳定位的误差理论，并为开发和部署大规模无线传感器网络及相关应用提供了重要的理论指导和实用方法。

内容获取失败，请点击重试