基于平面特征的LIO增强GNSS RTK定位性能研究  

Research on plane-feature based LIO enhancing GNSS RTK positioning performance

作　　者：孙之远 李星星[1] 汪世文 李圣雨 SUN Zhiyuan;LI Xingxing;WANG Shiwen;LI Shengyu(School of Geodesy and Geomatics,Wuhan University,Wuhan 430079,China)

机构地区：[1]武汉大学测绘学院,武汉430079

出　　处：《导航定位学报》2024年第1期12-20,共9页Journal of Navigation and Positioning

基　　金：国家重点研发计划项目(2021YFB2501102)。

摘　　要：针对应用单一传感器的智能导航定位技术在城市复杂环境下难以满足准确、无缝和稳定性能要求的问题,提出一种基于平面特征的雷达惯性里程计(LIO)增强全球卫星导航系统(GNSS)实时动态定位技术(RTK)定位性能的方法:通过滑动窗口构建多帧激光雷达(LiDAR)平面特征的关联;并将RTK与LIO在位置域层面直接融合;最后在开阔环境和城市环境中进行实验。结果表明,RTK/惯性导航系统(INS)/LiDAR组合系统能够在GNSS挑战环境中达到分米级精度;同时,在LiDAR平面特征的约束下,该方法速度和姿态的估计性能也可得到改善。Aiming at the problem that intelligent navigation and positioning technology using a stand-alone sensor is difficult to meet the requirements of accurate,seamless and stable performance in complex urban environments,the paper proposed a method of plane-feature based light detection and ranging(LiDAR)inertial odometry(LIO)enhancing global navigation satellite system(GNSS)real-time kinematic(RTK)positioning performance:the association of multi-frame LiDAR plane features was constructed through a sliding window;and RTK and LIO were directly fused at the position domain;finally,an experiment was conducted in the open-sky scenario and urban environments,respectively.Result showed that the RTK/INS/LiDAR integration system could achieve decimeter-level accuracy in GNSS-challenging environments;meanwhile,under the constraint of LiDAR plane features,the estimation performance of velocity and attitude of the proposed method would also be improved.

关 键 词：多传感器融合 实时动态差分定位技术(RTK) 激光雷达(LiDAR) 雷达惯性里程计(LIO) 惯性导航系统(INS) 

分 类 号：P228[天文地球—大地测量学与测量工程]