| 摘要: |
| 为了解决当前密闭环境下超宽带（Ultra-wide Band,UWB）技术在非视距（Non Line of Sight，NLOS）条件下无法实现精确定位的问题，提出了一种行人航迹推算（Pedestrian Dead Reckoning,PDR）辅助UWB结合移动基站融合定位方法。首先，将已知位置的标签作为移动基站，利用移动基站和固定基站与标签的通信进行距离测量以增加 LOS 情况的定位信息。其次，通过分析和利用PDR的高频噪声特性，能够在融合过程中有效抑制UWB的NLOS误差。提出改进的扩展卡尔曼滤波（Extended Kalman Filter,EKF)算法，通过对多传感器（UWB、PDR和移动基站）的信息进行更为精细的融合，来处理非线性问题。与单纯UWB/PDR数据融合算法以及使用KF融合算法相比，所提的融合策略定位误差分别提升0.6 m和1.8 m左右。 |
| 关键词: 室内定位 行人航迹推算 扩展卡尔曼滤波 |
| DOI：10.20079/j.issn.1001-893x.240626001 |
|
| 基金项目:山西省科技创新人才团队专项计划（202304051001035）；山西省科技成果转化引导专项（202204021301055）；山西省专利转化专项计划项目（202302003）；山西省留学回国人员科研项目（2021-133） |
|
| An Improved EKF Fusion Localization Method for NLOS Scenarios |
| SHEN Chenning,LIU Ning,YANG Cuiting,LIU Hong,WANG Jian揳n,LI Meiling |
| (1.Shanxi Communication Tongda Microwave Technology Co.,Ltd.，Taiyuan 030006,China;2.College of Electronic Information Engineering,Taiyuan University of Seience and Technology,Taiyuan 030024,China) |
| Abstract: |
| In order to solve the problem that ultra-wide band(UWB) technology cannot achieve accurate positioning under non-line-of-sight(NLOS) conditions in the current closed environment,a pedestrian dead reckoning(PDR) assisted UWB combined with mobile base station fusion localization method is proposed.Firstly,the tag with known position is used as a mobile base station,and the communication between the mobile base station and the fixed base station and the tag is used for distance measurement to increase the positioning information of LOS situation.Secondly,by analyzing and utilizing the high frequency noise characteristics of PDR,the NLOS error of UWB can be effectively suppressed in the fusion process.An improved extended Kalman filter(EKF) algorithm is proposed to deal with nonlinear problems by fusing information from multiple sensors(UWB,PDR and mobile base station) more precisely.Compared with that of the UWB/PDR data fusion algorithm and the KF fusion algorithm,the positioning error of the proposed fusion strategy is increased by about 0.6 m and 1.8 m，respectively. |
| Key words: indoor localization pedestrian dead reckoning extended Kalman filter |
