Coordinated Motion Control Of A Robotic Arm Based On Visual Servoing And Joint Positio

As one of the most popular robots,joint robots have the ability to simulate human joints and have increasingly promising applications in various fields.Therefore,there are more stringent requirements for mechanical arms in motion planning,grasping accuracy,etc.However,the working environment and grabbing targets of mechanical arms often have uncertainties.In order to solve the problems of spatial collision and target deviation of mechanical arms,this paper proposes a control scheme based on visual servo and joint position.The research contents are as follows.(1)Aiming at the problem of spatial collision of the mechanical arm in the operation process,a path planning algorithm based on improved Rapidly-exploring Random Trees(RRT)is designed.The algorithm uses a target-biased extension point selection strategy,introduces target deviation probability in the iterative process,and expands the tree in the direction of the target point when conditions are met,solving the high randomness and long convergence time of traditional RRT algorithm in path planning.The redundant nodes are selected by path optimization strategy,and the path is smoothed by cubic spline interpolation.The results show that the improved RRT path planning algorithm enables the mechanical arm system to reach the target position quickly and smoothly without collision.Compared with the traditional RRT algorithm,the path planning efficiency is improved by approximately 61.57%.(2)Aiming at the problem of target workpiece position deviation in the mechanical arm grasping task,a visual localization and coordinated motion control strategy is designed.After the mechanical arm reaches the target position without collision through path planning,the mechanical arm switches from joint position control to visual servo control through coordinated control strategy.Firstly,the target feature information is improved through image preprocessing,and the Canny operator is used to identify the contours in the environment.The Hough circle transform algorithm and the polygon fitting method are used for double detection of circular workpieces and trays,and the coordinates of the target object are obtained.The results show that the visual localization and coordinated control strategy has good positioning effects in the case of target position deviation,and the average position errors in the X and Y directions are 1.382 mm and 1.132 mm,which meet the requirements of mechanical arm grasping.(3)Aiming at the uncertainty of target object depth information,an improved Semiglobal Matching(SGM)stereo matching algorithm is designed.The cost calculation uses an improved Census transform,replacing the center pixel with the average of the neighborhood pixels,enhancing the robustness against noise.Based on multi-path cost aggregation,the aggregation penalty parameters are adaptively adjusted according to the gradient difference between the left and right views,improving the aggregation accuracy in low texture areas.Finally,the winner-takes-all algorithm and disparity optimization strategy are used to obtain the target disparity.The results show that in the case of unknown depth information,the algorithm can obtain high-precision,smooth and effective target disparity and depth,and the average height error in the Z direction is 0.708 mm,meeting the requirements of mechanical arm height positioning.The paper conducts trajectory planning from the perspective of space avoidance,allowing the mechanical arm to perform tasks in the workspace without collision;from the perspective of target recognition,a visual positioning and coordinated control strategy is designed to accurately grasp the target workpiece when the position changes;from the perspective of depth estimation,a stereo matching algorithm is designed to enhance the anti-interference ability and accuracy of depth information.The coordinated control strategy based on visual servo and joint position enables the system to have the fastness of reaching the target position without collision and the robustness of the target workpiece position,which is of great significance for industrial robot control.