Research On Environment Modeling And Exploration Of Minitype Unmanned Aerial Vehicles Based On RGB-D Data In Indoor Environments

Autonomous exploration mapping is a key technology for robots to realize intelligence.In the unknown environments,the environmental map is created after robots obtain the environment information by sensors.Meanwhile,robots need to determine the area which will be explored in the next moment and plan the path to that area according to the map.Autonomous exploration mapping includes three components:Simultaneous localization and mapping(SLAM),exploration,and path planning.When carrying out navigation task in the complex three-dimensional environments,the ground mobile robots have the problems such as low efficiency and narrow field of vision,etc.Compared with the ground mobile robots,minitype unmanned aerial vehicles(UAV)have the characteristics of high working efficiency and wide field of vision,etc.UAVs are widely used in search and rescue,aerial photography and power inspection,etc.UAVs are more suitable for the autonomous exploration mapping in complex environments.In this thesis,UAVs with depth camera is taken as the research object.Based on the information of red,green,blue and depth(RGB-D),RGB-D SLAM,environmental exploration and path planning are studied to realize the modeling and exploration of indoor environments.The main work of this thesis is as follows:Firstly,the problem of UAV RGB-D SLAM modeling is analyzed.The mapping method based on Octree Map(OctoMap)is studied.In order to improve the real-time of mapping,an optimization method based on submaps is proposed.The thesis analyzes the limitations of the conversion of point cloud to OctoMap and proposes a method based on point complement to construct OctoMap with complete map information.RGB-D SLAM is used to estimate the pose of UAVs and provides map information for environmental exploration.Secondly,the exploration method based on wavefront algorithm is proposed.The frontier points are searched according to the outward diffusion process of wavefront algorithm in OctoMap.When the frontier points are obtained,the evaluation function of frontier points is established based on the diffusion distance,the number of unknown voxels and the number of occupied voxels.The optimal frontier point is selected.In the exploration strategy,layering exploration strategy is adopted to improve the order of the whole process of autonomous exploration mapping and the flight stability of UAVs.Thirdly,the path planning method for UAVs integrated environmental exploration is proposed.The thesis analyzes the advantages and disadvantages of the existing path planning algorithms.Combined with the actual application of UAVs,the path planning problem of UAVs based on wavefront algorithm is studied,which is based on exploration framework.A smoothing method of path planning based on wavefront algorithm is proposed to reduce the number of path points and the number of turns.Finally,the simulation experimental platform and UAVs experimental platform are built.The simulation platform is used to verify the UAVs autonomous exploration mapping process,and the effectiveness of the proposed algorithm is verified by UAVs experiments in the actual environments.