摘要
视觉/惯性组合导航技术是自主导航领域的重要研究方向之一。首先介绍了视觉/惯性组合导航技术的发展概况,然后从纯视觉导航(里程计、同步定位与构图)以及组合导航(滤波、非线性优化)2个层次介绍了传统的基于视觉几何与运动学模型的视觉/惯性组合导航方法,还介绍了近年来发展迅猛的基于机器学习的视觉/惯性组合导航方法。最后,简要介绍了视觉/惯性组合导航技术的典型应用及未来发展趋势。
Visual-inertial integrated navigation technology is an important research direction in autonomous navigation.This paper first introduces the development of visual-inertial integrated navigation technology,then introduces the traditional visual-inertial integrated method based on visual geometry and kinematics model from the viewpoint of pure visual navigation(odometry,simultaneous localization and mapping)and integrated navigation(filtering,nonlinear optimization),then introduces the visual-inertial integrated navigation method based on machine learning which has developed rapidly in recent years.Finally,the typical applications and the future trend of visual-inertial integrated navigation technology are briefly introduced.
作者
张礼廉
屈豪
毛军
胡小平
ZHANG Li-lian;QU Hao;MAO Jun;HU Xiao-ping(College of Intelligence Science and Technology,National University of Defense Technology,Changsha 410073,China)
出处
《导航定位与授时》
2020年第4期50-63,共14页
Navigation Positioning and Timing
基金
国家自然科学基金面上项目(61773394)。
关键词
视觉里程计
同步定位与构图
惯性导航
组合导航
Visual odometry
Simultaneous localization and mapping
Inertial navigation
Integrated navigation
