复杂环境下微小飞行器惯性/激光雷达Robust-SLAM方法  

A Robust-SLAM Method for MAVs Using LIDAR and IMU in Complex Environment

作　　者：石鹏 赖际舟[1] 吕品[1] 邹冰倩 付相可 SHI Peng;LAI Ji-zhou;LYU Pin;ZOU Bing-qian;FU Xiang-ke(Nanjing University of Aeronautics and Astronautics,Nanjing211106,China)

机构地区：[1]南京航空航天大学,南京211106

出　　处：《导航定位与授时》2019年第1期14-21,共8页Navigation Positioning and Timing

基　　金：国家自然科学基金(61703207);江苏省自然科学基金(BK20170801);航空科学基金(2017ZC52017);江苏省六大人才高峰(2015-XXRJ-005);江苏省青蓝工程

摘　　要：针对微小型飞行器在巡检、探测和地图构建等应用中关键的自主导航技术,提出了一种基于惯性辅助的激光雷达Robust-SLAM方法用于微小型飞行器自主导航。相对于传统的激光雷达SLAM方法,该方法在SLAM框架中引入了感知环境突变检测方法,并且加强了惯性与SLAM的组合程度,有效地解决了高程方向感知环境发生突变时激光雷达SLAM定位误差大的问题。室内车库实际飞行实验结果表明,该方法能够实现微小型飞行器在三维空间中实时可靠的自主导航,具有较好的工程应用价值。Aiming at the key autonomous navigation technologies of MAVs(Micro aerial vehicles)in patrol inspection,detection and map construction,an autonomous navigation method for MAVs based on Robust-SLAM using a LIDAR with IMU(Inertial Measurement Unit) aided is proposed in this article.Compared to conventional LIDAR SLAM methods,an innovative detection function for detecting step changes in the explored environment is introduced in the framework of SLAM,and IMU and SLAM are integrated more tightly in the proposed method,making accurate realtime position estimation of MAVs available even in the case where exists dramatic changes in the perceive environment of LIDAR along the elevation direction.The performance of the proposed navigation approach in 3 Dspace is verified successfully on the autonomous navigation of a smallscale UAV with a real flight test in an indoor parking garage,and the results show that the proposed method has good potential in practical applications.

关 键 词：微小型飞行器 同步定位与构图 激光雷达 自主导航 

分 类 号：V249.32[航空宇航科学与技术—飞行器设计]