基于无迹卡尔曼滤波的移动目标跟踪算法  

Unscented KF-based Mobile Target Tracking Algorithm

作　　者：徐义晗[1,2] XU Yi-han(Huaian Vocational of Information Technology,Huaian 223003,China;Jiangsu Software Testing Engineering Technology Research and Development Center,Huaian 223003,China)

机构地区：[1]淮安信息职业技术学院,江苏淮安223003 [2]江苏省软件测试工程技术研究开发中心,江苏淮安223003

出　　处：《火力与指挥控制》2020年第12期149-152,共4页Fire Control & Command Control

基　　金：江苏省教育科学“十三五”规划课题(D/2018/03/46);淮安市创新服务能力建设计划基金资助项目(HAP201904)。

摘　　要：为了提高对目标跟踪的位置估计精度,提出基于无迹卡尔曼滤波的移动目标跟踪UKTT算法。UKTT算法先获取移动目标与锚节点间的接收信号强度指示RSSI值,再依据这些RSSI值训练通用回归神经网络GRNN模型,并由GRNN初步估计目标位置,再由无迹卡尔曼滤波UKF修正目标位置的估计值。仿真结果表明,提出的UKTT算法提高了对目标位置的估计精度。Target tracking is one of the typical applications of Wireless Sensor Networks(WSNs).Therefore,Unscented KF-based mobile Target Tracking(UKTT)algorithm is proposed in this paper.The UKTT algorithm first obtains the RSSI value between the moving target and the anchor node,and then trains the General Regression Neural Network(GRNN)model according to the RSSI value,and preliminarily estimates the target position by GRNN,and then revises the target position estimation by Unscented Kalman Filtering(UKF).Simulation results show that the proposed UKTT algorithm improves the accuracy of position estimation by combining GRNN and UKF.

关 键 词：移动目标跟踪 无迹卡尔曼滤波 回归神经网络 接收信号强度指示 

分 类 号：TPT393[自动化与计算机技术]