| With the development of robotics,the application of mobile robots in industrials,commercial and domestic use is increasingly common.Furthermore,people have higher requirements in intelligent robot.One of the most essential functions is that mobile robot can navigate in the map of surrounding environment and localize itself.Moreover,robot exploring the environment and creating the map is the foundation of navigation and localization.To avoid manual intervention,an autonomous approach based on contour of boundary is introduced in the dissertation.By using this method,a mobile robot equipped with low-cost 2D Lidar can explore the environment autonomously.To solve the global localization and kidnapping problem,the dissertation proposes a novel method by comparing the Lidar and map data and figuring out the matching score to calculate the robot pose without movement.Its result is the initial pose in robot dynamic tracking algorithm that is based on particle filter.The 2D Lidar can only detect the objects in a fixed plane.There exist obstacles,which cannot be detected by Lidar in environment.So the dissertation applies a method using sensor fusion by joint calibration.The obstacles in binocular camera is projected into Lidar point cloud.In navigation,the system would consider the Lidar and the cameras and plan a valid path to avoid the objects.For autonomous environment exploration,localization,navigation and sensor fusion,the dissertation illustrates a software system,which consists of those modules.The software has a server and a client.The client has interactive interface,and it realizes that user can click the function remotely,For example,user can point out a goal in map remotely and the robot go to the point avoiding the obstacles.Furthermore,in client,user can observe the robot state in real time.In experiments,the system,which has practical value,runs in high efficiency,and it is suitable for various conditions. |
PDF Full Text Request