Global Path-planning Method Under Air-Ground Cooperation

The autonomous navigation of autonomous vehicle in unknown complex environment has become a hot topic in the field of intelligent control.The sensor equipped by the autonomous vehicle has limited sensing ability for the large range of environmental information,and cannot obtain the global environmental information,which seriously restricts the ability of autonomous navigation of it.In this paper,we will build an air-ground coordination system for path planning.In this system,the UAV serves as a separable long-range vision system for the autonomous vehicle,providing top-down views of its environment,which are stitched and transformed into global maps.The system will make up for the lack of the ability of the mobile robot's perception of global environment,and will improve the capability of autonomous navigation of autonomous vehicle.In view of the effect of UAV's momentum on its real-time and accuracy of path navigation control,the state prediction based Unmanned Aerial Vehicle heading control method,a navigation control strategy,was presented to enable the UAV better follow the route set and to achieve more effective image acquisition about the global environment.In order to make use of information acquisition of image sequence to build the global environment,by using SIFT image stitching algorithm for image processing,and through the RANSAC algorithm for image matching error,and experiments show that this method can make the image stitching is more accurate and effective.Based on the above,we use K-means clustering segmentation algorithm and morphological algorithm to extract the road information of the global environment,the experiment shows that the method can achieve the accurate extraction of the road in the environment.Because the path planning is an important guarantee for the autonomous navigation of autonomous vehicle,it is necessary to study the path planning method for autonomous vehicle.The traditional path planning algorithm commonly present in local minima,smoothness and safety and other issues,this paper proposes a path planning method for autonomous vehicle based on Fast Marching Square algorithm.The result shows that the algorithm can effectively use the global environmental information collected by UAV and can obtain a global optimal path for the autonomous vehicle to track.The main characteristics and innovation of this thesis includes:1.Aiming at the limitation of in-vehicle sensor,we build an air-ground coordination system to map an area via UAV aerial imagery and aid an autonomous vehicle in navigating,to make up for the lack of the perception of global environment.2.A path planning algorithm for autonomous vehicle based on Fast Marching Square is proposed.3.We achieve the construction of a global environment through the image sequence mosaic,and use of clustering segmentation and morphology on the road to extraction.