基于磁航向约束的视觉惯性定位算法  

Visual/inertial positioning algorithm based on magnetic heading constraint

作　　者：李磊磊[1] 梁琳 陈家斌[1] 韩勇强[1] LI Leilei;LIANG Lin;CHEN Jiabin;HAN Yongqiang(School of Automation,Beijing Institute of Technology,Beijing 100081,China)

机构地区：[1]北京理工大学自动化学院,北京100081

出　　处：《中国惯性技术学报》2022年第2期188-194,共7页Journal of Chinese Inertial Technology

基　　金：国家自然科学基金(91420203);装备重大基础研究项目(51405-05B03)。

摘　　要：针对视觉惯性导航系统绝对航向信息缺失导致精度下降的问题,提出了一种基于磁航向约束的视觉惯性定位算法。首先,建立了视觉/惯性/地磁导航系统模型,并对惯性数据采用预积分进行处理,降低状态更新时系统计算负担,提高实时性;其次,在视觉惯性数据对齐后,引入地磁观测数据作为约束,获取真实航向信息,完成系统初始化;最后,构建目标函数,通过非线性优化实现视觉、惯性和地磁数据融合。通过搭建车载实验平台,在室外环境下对所提出算法进行了验证,实验结果表明,相比于VINS-mono,引入磁约束后,视觉惯性导航系统的平均定位均方根误差减小了约25%。Aiming at the lack of absolute heading information in the visual inertial navigation system leading to accuracy decreasing,a visual-inertial positioning algorithm based on magnetic heading constraints is proposed.Firstly,a visual-inertial-geomagnetic navigation system model is established.The inertial data is processed by pre-integration to simplify calculation during state update and improve real-time performance.Secondly,after aligning the visual inertial data,the geomagnetic observation data is introduced as a constraint to obtain the real heading information.Then the initialization of the navigation system is completed.Finally,the objective function is constructed by fusing the visual,inertial and geomagnetic data through the nonlinear optimization algorithm.The proposed algorithm is verified in the outdoor environment by building an on-board experimental platform.The experimental results show that compared with VINS-mono,the average root-mean-square error of the positioning in visual inertial navigation system with magnetic constraint is reduced by about 25%.

关 键 词：磁约束 同步定位与构图 非线性优化 紧耦合 闭环检测 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置]