基于激光SLAM辅助的陆用惯导行进间粗对准方法

A in-motion coarse alignment method for land inertial navigation system based on LiDAR SLAM

目前常见的陆用惯导系统的行进间对准方法往往存在一些各自的问题,如GPS辅助时容易受到信号遮蔽影响;而里程计辅助时往往会受到机动路线约束的限制。针对这些问题,提出了一种激光SLAM系统信息辅助的陆用惯性导航系统(SINS)行进间粗对准算法,推导了基于激光SLAM系统视在速度观测矢量的姿态矩阵更新算法,利用基于REQUEST的双矢量定姿方法,实现初始常值姿态矩阵C_(b)^(n)(0)的求取,完成行进间的姿态粗对准和定位导航。在实验室无人车实验中,所提方法的对准姿态误差比里程计辅助的方法减小了约56%;同时与合作单位的综合实验验证结果表明,在进行250 s的行进间粗对准时,最终时刻的航向角误差为-0.0088°,实现5 min内达到1 mil的粗对准精度,所提算法可以有效实现激光SLAM信息辅助下的陆用惯导的行进间粗对准,并具有较快的收敛速度。

At present,there are some problems in the in-motion coarse alignment methods of land inertial navigation systems,such as signal masking when assisted by GPS and being limited by maneuvering route constraints when assisted by odometer.In order to solve these problems,a in-motion coarse alignment algorithm of SINS based on LiDAR SLAM is proposed.The updating algorithm of the attitude matrix based on the observed vector of velocity of LiDAR SLAM system is derived.By using the double vector attitude determination method based on REQUEST,the initial constant attitude matrix C_(b)^(n)(0) is obtained,and the coarse alignment and position navigation are completed.In laboratory experiments with unmanned vehicle,the alignment error of the proposed method is reduced by about 56%compared with the odometer assisted method.Meanwhile the comprehensive experimental results of cooperation with other institutions show that the azimuth Angle error at the final moment is-0.0088°when the coarse alignment takes 250 s,achieving a rough alignment accuracy of 1 mil within 5 minutes,which is higher than that of the odometer assisted method under the same conditions.The proposed algorithm can effectively realize the in-motion coarse alignment of the inertial navigation system with the assistance of LiDAR SLAM that has a fast convergence rate.