Environment Perception And Intelligent Control Of Loading Robots

The logistics industry has grown at an unparalleled rate in recent years,thanks to the advent of the home network economy.As a result,items in the marketplace must be disseminated fast and in big quantities.Various industrial automation technologies have begun to appear in the loading and handling process as the contradiction between the quick rise in demand for goods handling and the low efficiency of manual loading and handling becomes more and more problematic.Due to the limited applicability of the semi-automatic loading method based on teaching and reproduction of the currently developed box-type cargo loading robot,the box-type cargo intelligent loading system’s three-dimensional vision measurement scheme,the identification and positioning of the cargo box’s internal environment,and the loading robot The three features of autonomous motion planning were investigated in depth,and a loading robot closed-loop control system based on visual feedback was built.First and foremost,by examining the issues of low cargo loading efficiency,excessive labor workload,and low intelligence of present loading robots in the logistics business.Instead of using the human eye recognition function,the new TOF three-dimensional camera is put at the end of the loading robot to achieve intelligent detection of the truck compartment’s environmental information.The hand-eye system,which consists of the camera and the loading robot,is calibrated so that the extracted coordinates from the 3D camera can be transferred to the loading robot’s base frame and exact positioning may be achieved.Second,a statistical filtering technique based on the point cloud depth information constraint is proposed,based on the features of the data structure of the point cloud captured by the 3D camera,in light of the problem that the existing point cloud filtering effect is not visible.Then,to eliminate the mixed points in the point cloud data,apply the least squares approach to smooth the data.A slicing method is proposed based on the point cloud information characteristics of the cargo wall inside the cargo compartment to remove significant background point cloud information from the point cloud data,and then use the region growing algorithm to segment the point cloud data and extract the cargo point cloud data.Finally,for forward and inverse kinematics analysis,a kinematics model is created based on the loading robot.After loading and positioning the collected point cloud data,the path points are interpolated in the robot joint space.Realize autonomous loading robot planning and control.The data accuracy of the target point cloud for visual extraction and the actual effect of the vision-guided loading robot for loading items were examined using an experimental platform for intelligent loading.This paper investigates the challenges of the logistics industry’s rapid growth,as well as the lack of spatial perception and intelligent control of loading robots,and proposes a new TOF three-dimensional camera-based spatial perception and intelligent control method for loading robots that replicates the original teaching and controls.The visual feedback control method replaces the approach,which considerably increases the loading robot’s external space perception capacity and real-time decision-making ability,and has good engineering importance and application value.