室内退化场景下UWB双基站辅助LiDAR里程计的定位方法

Research on localization method of UWB assisted LiDAR odometry in degradation scene

针对在结构特征稀疏且连续重复的退化场景下,基于激光雷达(LiDAR)的同时定位与建图技术(SLAM),由于约束缺失而无法恢复运动状态的问题,笔者提出利用超宽带(UWB)双基站辅助LiDAR里程计的定位方法。首先,利用退化因子判断LiDAR SLAM的退化方向及退化程度;然后,利用UWB测距值提供距离约束,并结合LiDAR里程计信息以构建状态方程和量测方程;最后,通过扩展卡尔曼滤波进行参数解算。实验结果表明,笔者提出的方法能够在室内退化场景下获得准确且低漂移定位结果,定位精度为亚米级,相较于LiDAR定位与地图构建(A-LOAM)算法,其均方根误差降低了22.3%,平均误差降低了21.23%。

Aiming at the problem that the Simultaneous Localization and Mapping based on Light Detection and Ranging(LiDAR SLAM)can not recover its motion state due to the lack of constraints in the degraded scene with sparse and continuous structural features,a location method of LiDAR odometry assisted by Ultra-Wide Band(UWB)dual base stations is proposed in this paper.Firstly,the degradation factor is used to judge the degradation direction and degree of LiDAR SLAM,then the UWB ranging value is used to provide distance constraints,and the LiDAR odometry information is combined to construct the state equation and measurement equation.Finally,the parameters are solved by extended Kalman filter.The experimental results show that the proposed method can obtain accurate and low drift positioning results in indoor degraded scenes.The positioning accuracy is sub meter level.Compared with a Lidar Odometry and Mapping(LOAM)algorithm,the root mean square error is reduced by 22.3%and the average error is reduced by 21.23%.