Motion Control Of Automatic Feeding Manipulator Based On Machien Vision

Machinery manufacturing automation has been the goal of manufacturing industry for thesake of long-term development. The implementation of automation can greatly reduce the laborintensity, improve product quality, enhance the ability of enterprises to adapt to the market, andimprove competitiveness. In this paper, the motion control of automatic feeding manipulator wasresearched. The main research contents are as follows:(1) Design of the automatic loading control system: The system includes a transmittingdevice AGV, robots, material and visual system, and all the parts cooperate with each otherthrough the control. AGV transported the box filled up with materials, machine vision was used toget images of the parts inside the box, and then the parts posture obtained by image processingwas sent to the robots, so that the manipulator was drived to grab parts and sent them to thedesignated location.(2) Kinematics analysis of the manipulator: the D-H method is adopted for manipulatorkinematics analysis, the D-H coordinate system was established, D-H parameters were analyzed,the manipulator kinematics model is established. The positive solutions of are derived, usingmovement characteristics of the manipulator and piper criterion for manipulator inverse solutionto get joint variables.(3) Vision-based gesture detection of parts: the camera calibration, image acquisition, imageprocessing, identification features, and ultimately determine the attitude of parts. Zhang’scalibration method was employed to calibrate the CCD camera, a single part was filmed, and thenthe part template acquired by image processing was saved to identify the parts of the visual systemafter feedback. Through the image processing software-Halcon to do real-time image acquisition,image preprocessing, and then using a template matching method to get the gesture s of variouspart in order to convenient the robot grasp.(4) Robot motion control: kinematic analysis was adopted to obtain a positive solution andinverse solution model, trajectory planning algorithm and process was designed, using matlab todo data planning and simulation, a series of points was obtained, and trajectory planning curvewas drew. Robot three-dimensional modeling was done in solidworks, and then import the discretedata acquired by matlab into solidworks for Cosmos simulation analysis, also got the trajectoryplanning curve. The algorithm was verified through the comparison of the two trajectory planningcurves in solidworks and matlab.