The Research On Navigation And Visual Capture Method Of Omnidirectional Wheel AGV

Automated Guided Vehicle(AGV)is one of the core equipments of modern workshop logistics system,and it has huge market space.On the one hand,the existing traditional mobile robots on the market do not have independent operation and sorting functions.On the other hand,the purely robotic handling robots can only achieve actions such as grasping,handling and assembling within a fixed range Limited working space.The omnidirectional wheel AGV is used as the moving platform and the manipulator is carried on.It navigates along the magnetic strip to the platform to load and unload materials automatically.Research on navigation and vision grasping method of AGV with omni-directional wheel.The PD algorithm compensates in the RFID localization and the magnetic navigation to carry on the calibration to the vehicle body localization and the movement path accuracy,and carries on the navigation localization experiment,enhances the movement accuracy of the movement navigation;The image of the material is preprocessed,and the image dedistortion and Hu Matrix are used to solve the problems of classification and adhesion,and the grasping test is carried out to improve the Suction Cup position and grasp the material accurately According to the function requirement of AGV,the whole scheme of omni-directional wheel AGV is designed,and the design and realization of AGV movement and navigation system are described in detail.In order to solve the movement problem of AGV,which is not flexible enough to work in the limited working condition,the AGV is loaded with mcnamara Omnidirectional wheels to solve the problem of AGV moving to the working position in the interior environment such as workshop The magnetic stripe navigation scheme of AGV and the combination strategy of site identification technology are studied,the site identification method is designed by using magnetic sensors which are necessary for magnetic stripe navigation According to the mechanical relation of the manipulator,the kinematics analysis of the manipulator is carried out to verify the correctness of the model A visual image capturing algorithm based on OPENCV library is designed to recognize the grasping point and estimate the grasping position and pose of the grasping point.An obstacle detection method based on SSD-300 is studied.In the unknown environment,whether the AGV is continuously moving,loading and unloading or stationary,MJPG video stream coding and network transmission are used The remote maintenance personnel know the video stream information of the surrounding environment of AGV in time(by receiving the alarm information of finding obstacles through the mobile phone),and design an image and video acquisition software with Ui based on PyQt5 It solves the problem of AGV stopping on the magnetic strip which is caused by the obstacle that blocks AGV's movement outside the experiment time.Finally,the average accuracy of the detection is 0.89,and the experimental results verify the effectiveness of the design and Algorithm.