Research On Path Planning Of Ground Robot Based On UAV's Vision

In recent years,due to the ability of a single robot reaching bottleneck,the research of multiple robot is increasingly,and is applied to the target search,rescue,surveillance and some other fields.However due to the complexity of the collaboration of unmanned aerial vehicles(UAVs)and a ground robot,there isn't many existing researches of ground robots'path planning based on UAV visual.The existing research mainly concentrates in the rescue scene.The UAV's open vision and mobile advantage help the ground robot in the capacity of exploring environment so that improving the ability of ground robot's navigation in complex environment and task execution.This paper research on some related problems of the ground robot path planning based on UAV visual,mainly about environment modeling based on UAV visual and path planning technology,and so on.And depending on whether real-time acquisition environmental information,the research is divided into the global path planning and the local path planning scheme.Firstly,the conversion relationship between camera coordinate system and the world coordinate system is studied so that the environmental information obtained from the camera can be converted.The method of the image mosaic and the detection of obstacles are studied,and these methods are verified in the current experimental environment.Secondly,this paper studies the environment modeling and path planning algorithm in the global map.In this paper,the three methods are studied in detail,namely the path planning algorithm based on the Voronoi diagram,path planning based on the a*algorithm in grid environment and artificial potential field method.And then this paper shows simulation experiment analysis,verifying the validity of these methods.Finally,a new local path planning method combination of rolling window method and the a*algorithm is proposed.In the scene of environment information partially known,a more effective local path planning algorithm is proposed.