Target-based Rescue Robot Autonomous Navigation And Manipulator Motion Planning

The rescue environment usually has extremely high uncertainty,complexity and danger.The main function of the rescue robot is to replace the rescuers to complete the search and rescue mission.This task can be divided into two parts,one is to complete the autonomous exploration of the environment and determine the location of the victim,and the second is to use the robotic manipulator to carry out further rescue operations.In this thesis,the target-based rescue robot autonomous navigation and manipulator motion planning are proposed and implemented in the environment simulated by the RoboCup rescue competition.For the first part of the task,this thesis combines the autonomous exploration of robots in the rescue environment with target recognition.We propose a target-based robot autonomous exploration method.In this method,once the victim is found,the robot will use the victim as the navigation target,use the timing elastic band method based on the graph optimization to directly perform the trajectory planning,and quickly reach the area of the victim to proceed to the next rescue operation.This method makes full use of the target information to assist navigation,and makes up for the defects of large randomness and slow navigation speed of the frontier-based exploration method,which greatly improves the exploration efficiency of the robot in the rescue environment.For the second part of the task,this thesis combines the target recognition and localization algorithm with the motion planning algorithm of the manipulator.Using the UR3 manipulator and the Kinect v2 depth camera to form the hand-eye coordination system,the coordinate transformation relationship between the Kinect camera and the UR3 manipulator is first determined.Therefore,the pose of the target in the robot arm coordinate system is obtained.And then a trajectory planning method based on linear interpolation is proposed.Compared with the sampling-based trajectory planning method,the planned path is optimal.Moreover,the experimental results show that the motion planning method based on linear interpolation has shorter trajectory planning and execution time.