NLOS环境下多传感器融合的室内定位方法  

Indoor positioning method of multi-sensor fusion in NLOS environment

作　　者：张倩榕 李毅[2] ZHANG Qian-rong;LI Yi(National Key Laboratory of Fundamental Science on Synthetic Vision,Sichuan University,Chengdu 610065,China;College of Computer Science,Sichuan University,Chengdu 610065,China)

机构地区：[1]四川大学视觉合成图形图像技术国防重点学科实验室,四川成都610065 [2]四川大学计算机学院,四川成都610065

出　　处：《计算机工程与设计》2023年第3期732-738,共7页Computer Engineering and Design

基　　金：科技部国家重点研发计划基金项目(2018YFB1601201-3)。

摘　　要：为提高超宽带(ultra-wideband,UWB)技术在非视距(non line of sight,NLOS)环境下的定位精度,提出一种基于粒子滤波融合视觉与UWB数据的定位方法。视觉模块通过识别与检测标签推算出绝对位姿;UWB模块鉴别由NLOS条件干扰的测距值,筛选最优测距值进行自适应权重的定位算法,提升覆盖区域的整体定位精度;基于粒子滤波将两者的实时定位信息进行数据融合。实验结果表明,融合定位方法具有实时性和鲁棒性,有效抑制了NLOS环境引起的误差,在NLOS环境下定位精度能够达到0.26 m。To improve the positioning accuracy of ultra-wideband(UWB)technology in non-line of sight(NLOS)environment,a positioning method based on particle filter fusion of vision and UWB data was proposed.The visual module calculated the absolute position and pose by identifying and detecting landmarks.The UWB module identified the ranging values interfered by NLOS conditions and selected the optimal ranging values for the localization algorithm of adaptive weight,which improved the overall positioning accuracy of the covered area.The real-time location information was fused by particle filter.Experimental results show that the fusion localization method is real-time and robust,and can effectively suppress the errors caused by NLOS environment,and the localization accuracy can reach 0.26 m in NLOS environment.

关 键 词：多传感器融合 室内定位 粒子滤波 超宽带 视觉定位 非视距 移动机器人 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]