Feedback-based Target Localization in Underwater Sensor Networks: A Multi-Sensor Fusion Approach
Feedback-based Target Localization in Underwater Sensor Networks: A Multi-Sensor Fusion Approach
复制标题
水下传感器网络中基于反馈的目标定位：多传感器融合方法
DOI：
10.1109/tsipn.2018.2866335
复制
发表时间：
--
影响因子：
3.2
通讯作者：
Xinping Guan
中科院分区：
文献类型：
--
作者：
Jing Yan;Ziqiang Xu;Xiaoyuan Luo;Cailian Chen;Xinping Guan
本文研究了水下传感器网络中目标定位问题，受到测量范围和精度的限制。本地化过程主要分为两个阶段，即，距离估计和位置求解。在第一阶段，由于水下环境的高动态和强噪声特性，部分传感器节点无法获得目标的直接距离测量值。在此基础上，将距离测量问题转化为闭环控制问题，设计了一种比例积分估计器，通过间接测量获得传感器的距离信息。利用估计的距离信息，在第二阶段提出了一种基于一致性的无迹卡尔曼滤波（UKF）算法来定位目标，其中融合了直接和间接测量以减少恶意数据的影响。此外，稳定性条件表明，距离估计器可以稳定的闭环系统，而有界性分析证明，以保证定位精度。仿真结果表明，与单一的直接测量相比，该距离估计器可以扩展传感器的测量范围。同时，基于一致性的UKF算法可以有效提高定位精度。
This paper investigates the problem of target localization in underwater sensor networks, subjected to limited measurement range and accuracy. The localization process is mainly divided into two phases, i.e., distance estimation and position solving. In the first phase, some sensor nodes cannot acquire the direct distance measurements of target, due to the high-dynamic and strong-noise characteristics of underwater environment. Based on this, we formulate the distance measurement as a closed-loop control problem, and then a proportional-integral estimator is designed for sensors to acquire the distance information through indirect measurements. With the estimated distance information, a consensus-based unscented Kalman filtering (UKF) algorithm is proposed in the second phase to localize the target, where direct and indirect measurements are fused to reduce the influence of malicious data. Moreover, stability conditions are provided to show that the distance estimator can stabilize the closed-loop system, while the boundedness analyses are demonstrated to guarantee the localization accuracy. Finally, simulation results reveal that the proposed distance estimator can extend the measurement range of sensors by comparing with the single direct measurement. Meanwhile, the consensus-based UKF algorithm can effectively improve the localization accuracy.