| In recent years,the field of robotics has developed rapidly around the world.Among them,mobile robots,as a master of robotics technology,have received increasing attention from experts and academics all over the world.Mobile robots need to have good mobility and excellent autonomous operation capabilities.However,at present,most mobile robots have problems such as low positioning accuracy and the need for professional operators to assist in their operations,which severely restrict their development.Therefore,how to increase the positioning accurateness of mobile robotic arm and realize their fully autonomous operations is a popular research theme.This paper takes the mobile service robot for helping the elderly and the disabled as the design object,and researches on the background of taking and delivering objects.It focuses on the navigation and positioning of the mobile platform based on the QR code,the robotic arm grasping based on point cloud information,and the man-machine based on the user's joint gestures.Specific research has been done on these key technologies of interaction,and the main research contents are as followsFirstly,the research status of mobile robot arm,robot arm gesture control,and 3D vision-based robot arm grasping technology are introduced,and then the overall scheme of the mobile manipulator studied in this paper was designed,the vision system and control system of the mobile manipulator were built,and the kinematics analysis of the mobile manipulator was carried out.Aiming at the problem that the laser positioning of the mobile platform is not accurate,which will cause the target object to not be in the working space of the robotic arm,this article uses a QR code reader to identify the landmarks of the QR code for navigation and positioning.The position accuracy can reach 10.2mm,and the angle accuracy can reach ±0.1°.It ensures the reliability and robustness of the positioning of the mobile robotic ann,which is very suitable for navigation and positioning in a home environment with a fixed route,and provides a foundation for the subsequent robotic arm to successfully complete the grasp.In view of the fact that the traditional two-dimensional vision-based robotic arm cannot adapt to the unstructured working environment,which greatly affects the scope of use of the mobile robotic arm,this paper uses the Real Sense to identify the object to be grasped,and realizes accurate grasping of objects in different poses and different levels.First,use the hand-eye relationship of "eye-in-hand" to calibrate the robot arm;then perform point cloud preprocessing,plane detection and filtering,object segmentation,and template matching based on ICP algorithm to obtain the grasping posture;At last,the route planning and trajectory planning of the robotic arm are carried out to realize the autonomous grasping of the target object.Aiming at the fact that mobile robotic arms cannot achieve fully autonomous operations and their operations are complex,its operators often need to be specially trained,and are not suitable for the elderly and disabled person in the home environment.In this paper,the Kinect depth camera is used to obtain the user's joint and gesture information to realize the remote control of the robotic arm,making the human-computer interaction efficient and natural,and greatly improving the operating efficiency of the robotic arm.Finally,the positioning experiment,object grasping experiment and human-computer interaction experiment of the mobile platform are carried out.The experimental results show that the RGB-D-based mobile robotic arm grasping system and human-computer interaction system studied in this paper have good performance,good accuracy and reliability,and can achieve a certain degree of autonomous operation in the family's non-structural environment. |
PDF Full Text Request