Research On Flexible Grasping Teaching Method Of Mobile Manipulator

Mobile manipulator has gradually entered people's field of vision,can replace people to perform some tasks.When the mobile manipulator carries out the task of mobile grasping,it is often restricted by environmental factors,such as more obstacles,more complex environment,and the presence of occlusion of the captured objects,which leads to the low success rate of the mobile manipulator's grasping planning and the long planning time.However,the traditional teaching method of manipulator grasping is only suitable for grasping in fixed place,not suitable for the changing environment of mobile manipulator.Aiming at the limitations of traditional teaching method,this paper proposes a flexible grasping teaching method of mobile manipulator.The main research content of this paper is how to optimize the docking position of the mobile manipulator and combine with the grasping mode of teaching,so as to improve the planning efficiency of the manipulator grasping action,including improving the success rate of planning and reducing the planning time.In this paper,firstly,the workspace of the manipulator is analyzed,and the evaluation index of grasping difficulty of the manipulator is obtained by randomly scattering points,and the position of the target object,obstacle and mobile platform in the environment is simplified as a point and projected on the XY plane.According to the evaluation index of grasping difficulty,the docking position of the mobile platform is finally obtained;Aiming at the obstacle avoidance problem of the manipulator,RRT algorithm is used to realize the obstacle avoidance of the end of the manipulator,the link of the manipulator and obstacles,and the collision between the end of the manipulator and its own link.The simulation results show that the success rate of grasping objects increases from 76.53% to 99.32%,and the obstacle avoidance planning time of RRT algorithm is reduced by 44.46%.In order to make the prototype experiment more in line with the actual environment,first of all,under the square table and the long table,the grab experiment is built according to the simulation environment;In addition,the motion capture system is added to carry out the experiment of real-time capture according to the location information in the environment;Finally,considering the error of the mobile platform,the realsense camera is added to achieve the secondary positioning of the object to ensure the accuracy of grasping the object.The above experiments prove the effectiveness of the method.