Research On Position Control Of Robot Arm Based On Visual Feedback

Robots have already been playing a significant role in human production and life since their inception.As the most widely used robot,robotic arm has gradually become a research hotspot.In order to improve the intelligent level of robotic arm,the application of vision sensor has received extensive attention.This paper introduced the application history and the present situation of vision technology in robotic arms,and analyzed the traditional method of realizing the position control of the robotic arm which based on kinematics.According to the traditional position control mode based on kinematics,this paper simulated the way that people control their arm to grab the target object with the help of vision and proposed a corresponding position control which is based on visual feedback.This method is suitable for robotic arms with three degrees of freedom or less.Based on the position relationship between the target point and the robotic arm on a visual plane,the movement direction of the joint can be determined by this method,and the movement speed of the joint can be determined by fuzzy control method,so that the end of the robotic arm can be controlled to reach the desired position.According to the control idea,the overall plan was completed firstly.By obtaining the coordinates of the target point,the end and the joints of the robotic arm on a visual plane,the mathematical model of each robotic arm based on visual feedback was established,and the algorithm which can judge the direction of joints was designed secondly.Finally,a fuzzy controller that can meet the control requirements was built,so that the position control of the robotic arm can be realized.To prove the effectiveness of the method,A virtual simulation environment and the three-dimensional visual model of each robotic arm were built in V-rep firstly.Secondly,the communication links between V-rep and the Matlab were established,and the co-simulation could be realized.During the simulation,the markers were detected with an ellipse detection algorithm based on arc-support line segments and a circle detection algorithm based on Hough transformation,so that the coordinate information was obtained.Finally,the simulation experiments of position control for each robotic arm were carried out.To further prove the effectiveness of the method,an experimental platform of two-bar linkage was built,and experiments were carried out.USB camera was called to capture visual images,and the image processing module was built in Labview,so that the markers could be tracked and the required coordinate information could be provided for the algorithm.Through the mixed programming of Matlab and Labview,the judgement algorithm for the two-bar linkage and the fuzzy controller were called,the speed and direction of the two motors were output.Through serial communication,the control commands were sent to motors to control the end of the robotic arm to move to the specified position.