Research On Robotic Arm To PC Components Steel Binding Technology With Visual Collaboration

With the rapid development of China’s assembled reinforced concrete building forms,precast concrete is widely used in the construction industry,but the reinforcement tying process in precast reinforced concrete still relies mainly on manual labor,which increases labor costs due to the increasing degree of aging,while the process requires workers to bend over for a long time,which is labor-intensive and has low productivity and cannot guarantee the tying quality.The construction of rebar tying system platform and the realization of smart tying are of great significance for smart construction.In the process of rebar tying,the robot is required to be able to quickly detect the status of the tying point and locate the coordinates of the untied point,and has good motion planning ability to quickly complete the task of rebar tying.In view of this,this paper builds a rebar tying robot system with depth perception capability for PC components,and investigates the robot arm autonomous rebar tying method in laboratory environment.Based on the improved YOLOX-s algorithm to detect the state of the rebar tying point;the rebar skeletonization and Hoff linear positioning rebar tying point methods are proposed;finally,the motion planning by RRT-Connect controls the robotic arm to move to the point to be tied to complete the tying task.The main research works are as follows.(1)Build the overall system of rebar tying robot.The hardware system of the robot arm is composed of AUBO-i5 robot arm,end-tying actuator and Kinect v2 depth camera,and the software system of the tying robot is composed of robot operating system(ROS)platform combined with vision and control algorithms.(2)Steel binding points state detection and localization.Firstly,the camera calibration technique and image alignment method are used to obtain the camera internal reference matrix,distortion parameters and alignment matrix of depth image and color image;secondly,the state of the rebar tie point is detected based on the improved YOLOX-s algorithm,and the pixel coordinates of the rebar tie point are located using the rebar skeletonization and Hoff linear detection algorithm;finally,the point cloud of the rebar image is obtained by combining the aligned depth image value and color image value to locate the rebar tie point Finally,we obtain the point cloud of rebar images by combining the aligned depth image values and color image values,and locate the coordinates of the rebar tying points in 3D under the camera coordinate system.(3)ROS-based control method for robotic arm hand-eye tying.Firstly,the AUBOi5 robot arm is modeled by D-H parameter method,the positive and negative kinematics of the robot arm are analyzed,and the accuracy of the established kinematic equations is verified in the MATLAB robotics toolbox;secondly,the URDF modeling of the robot arm is carried out in the Moveit! Connect algorithm to complete the motion planning of the lashing points;finally,we complete the hand-eye calibration based on Moveit!(4)Experimental study of rebar tying.The results show that the coordinates of the tying points identified and located are within the tying range of the tying execution device;the average time to locate the tying points is about 108 ms,and the average time to complete a tying is about 15 s,realizing the real-time tying requirement.