Research On SLAM For Ground-Aerial Robots Cooperation And Construct The Experimental Platform

Unmanned system autonomously navigate in unknown environments by GPS normally.However,GPS information is unreliable or even unmeasurable in some complex environments.Therefore,unmanned system require other sensors to measure or estimate its poses in GPS denied environments.Besides,a heterogeneous robot system formed by a ground robot and an aerial robot cooperation can achieve more complex missions or have higher task effectiveness.Thus,pose estimation problem with unmanned system and heterogeneous unmanned system cooperation problem possess important application value and research signification.In the above context,this paper research on cooperative motion control technology for small-sized UAV and UGV without GNSS information.This paper focus on the four aspects of the problems:(1)Pose and velocity estimation based on airborne monocular vision for UAV.(2)Mapping in unknown environments.(3)Cooperative control based on UGV navigation for UAV.(4)Data communication between UAV and UGV.Finally,an experimental platform is built,and the related algorithms were implemented to verify the key technologies.Aiming at the problem of pose estimation problem for rotorcraft,the algorithms and strategies involved in the implementation of mono-SLAM system have been studied,SLAM system scheme has been designed.Aiming at the problem of mapping in unknown environments,two kinds of dense map schemes have been designed,including point cloud map and octomap.A tracking control protocol has been designed,a controller has been built,control gains have been tuned and the execution strategy of the control protocol has been given.The performance of the control protocol has been verified by reliable navigation information provided by motion capture system.Aiming at the problem of data communication between UAV and UGV,a communication system based on ROS has been designed,and the real-time data interaction performance has been verified.Finally,a low-cost experimental platform for ground-aerial robot cooperative motion has been constructed.Based on that,SLAM-based indoor position experiments with UAV and UGV have been executed,indoor hovering and tracking experiments with UAV have been executed.The experiments above verify the reliability and instantaneity for navigation information,maps have been presented in indoor unknown environments,and the position cooperation experiment of the ground-aerial robot has been completed and analyzed.