基于KF-BA-LSSVR的无线传感器网络节点定位研究  

Research on Node Localization of Wireless Sensor Network Based on KF-BA-LSSVR

作　　者：束仁义 朱家兵 沈晓波 蔡俊 夏泐 SHU Renyi;ZHU Jiabing;SHEN Xiaobo;CAI Jun;XIA Le(School of Electronic Engineering,Huainan Normal University,Huainan Anhui 232038,China)

机构地区：[1]淮南师范学院电子工程学院,安徽淮南232038

出　　处：《重庆科技学院学报（自然科学版）》2022年第3期57-61,共5页Journal of Chongqing University of Science and Technology：Natural Sciences Edition

基　　金：安徽省科技重大专项“单片BDS GPS GALILEO多模导航SOC芯片研发与应用”(202003A05020031);安徽省重点研发计划“基于声波阵列多频探测的智能火灾预警设备关键技术及产品研发”(202004A05020023);淮南市指导性科技计划项目“LOS NLOS环境下基于TOF测距的超宽带室内定位技术研究”(2021096);淮南师范学院科研重点项目“基于UWB的矿井巷道移动目标定位系统设计与实现”(2020XJZD001)。

摘　　要：为进一步提高基于最小二乘支持向量回归机算法(LSSVR)的无线传感器网络定位精度,提出一种基于蝙蝠算法(BA)的LSSVR定位方法。考虑到目标定位受环境干扰,采用卡尔曼滤波算法来减小测距误差。仿真实验结果表明,在三维空间中,相较于传统LSSVR算法的无线传感器网络定位,经过BA算法优化后,节点的平均定位误差减小,在考虑高斯噪声的影响下,通过卡尔曼滤波进行距离优化后的定位精度得到明显改善。In order to further improve the positioning accuracy of the wireless sensor network based on least squares support vector regression(LSSVR)algorithm,a positioning method of LSSVR based on bat algorithm(BA)optimization is proposed.Considering the influence of environment disturbance in the target location,Kalman filter algorithm is used to reduce the ranging error.The simulation experiment results show that,compared with the traditional LSSVR algorithm,the average positioning error of the nodes optimized by BA algorithm is reduced in three-dimensional space,and the positioning accuracy of the nodes optimized by Kalman filter under the influence of Gaussian noise is significantly improved.

关 键 词：最小二乘支持向量回归机 蝙蝠算法 卡尔曼滤波 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]