Research On Accurate Positioning Technology Of Loading And Unloading System Of Mobile Robotic For Machine Tool

With the concept of industrial 4.0,industrial robot is widely used in all walks of life in recent years and require higher integration,intelligent.Iudustrial robot combined all kinds of mobile vehicle platform and vision to integration different sensors to adapt to all kinds of uncentain working envirments a research focus.The autonomous positioning and navigation of the mobile robotic arm can flexibly dispatch industrial robots to designated work sites while avoiding obstacles or adapting to multi-robot work scenarios.Among them,the vision system is used to solve the problem of object recognition and positioning of parts loading and unloading or assembly.Aiming at the single working path of the mobile trolley,it has poor adaptability and low flexibility in complex scenarios such as human interference or multiple robots.In this paper,a self-guided navigation based on a laser sensor is studied to move the six-axis robot to a specified position while avoiding dynamic obstacles.For the problem of accumulating errors when the mobile robot arm is repeatedly positioned on the loading and unloading,the research focuses on the fast and accurate positioning of the jig on the m achine tool by the industrial camera mounted on the robot arm,and places the workpiece on the jig of the machine tool.A precise positioning system based on compound robot mobile fusion monocular vision for loading and unloading on the machine tool was developed.The main work is as follows:Firstly,design and build the hardware and software architecture of the mobile robotic arm system with visual positioning system.Secondly,Using the encoder on the car to roughly locate the car,the laser radar and triaxial accelerator gyro data are extracted in real time,and the known data is fused.Compared with the Hector SLAM algorithm,the initial value selection has a great influence on the car positioning.In this paper,an efficient Rao-Blackwellized particle filter is used to update the map by re-sampling and updating the particles.The GMapping algorithm creates a grid map.In order to provide autonomous navigation for subsequent mobile robots,a global optimal path is planned.Thirdly,On the above map,we test and compare two popular global optimal path planning algorithms,Dijkstra and A *.Integrate the car with the motor,lidar and gyroscope data using the ROS platform.To achieve the mobile car navigate between pick-up point in the workpiece and the point of loading and uploading.In the process of fixed-point navigation,a real-time planning algorithm of dynamic sliding window is set up to realize the local path planning of the car and avoid obstacles.Finally,statistic and ananlyse the error of mobile car's fixed-point navigation.Through the fixed-point navigation experiment,the maximum statistical error of mobile cart positioning error is 29.704 mm.Fourthly,Establish a camera model with the camera attached to the end of the six-axis robot.Focusing on the Levenberg-Marquardt nonlinear optimization method with constrained constraints,the high-precision fitting of the eye relationship is performed.In the experiment,more than 14 sets of robot end tools and camera pose data were extracted,and the above-mentioned L-M algorithm was used to match the eye relationship.The accuracy of hand-eye calibration was within ±0.423 mm by experimental analysis.Fifthly,pick up the workpiece through the workpiece image recognition and positioning,the car fixed-point navigation to the workpiece starting position,get the camera position in the custom coordinate system,teach the workpiece into the machine tool and read the robot end tool position,extract more The secondary data solves the custom coordinate system and fixture relationship matrix.Finally,design mobile robot's loading and unloading experiment.Integration of industrial robots to complete the workpiece positioning and picking.Mobile car navigating to the initial position of the workpiece loading and unloading.Then,Rapid precise positioning of the clamp and drive industrial robot to work.The experiment ensure accurate secondary positioning within ± 0.5mm,the attitude angle within ± 1.5 °.