![[Tutorial] building machine learning models for predictive maintenance applic...](https://cdn.slidesharecdn.com/ss_thumbnails/tutorialbuildingmachinelearningmodelsforpredictivemaintenanceapplications-yanzhang-150820102412-lva1-app6891-thumbnail.jpg?width=560&fit=bounds)
More Related Content
Similar to DEVELOPMENT OF AUTOMATIC TEACHING METHOD USING STEREO CAMERA FOR SCARA ROBOTS (20)
Recently uploaded (20)
DEVELOPMENT OF AUTOMATIC TEACHING METHOD USING STEREO CAMERA FOR SCARA ROBOTS
- 1. VIETNAM NATIONAL UNIVERSITY HO CHI MINH CITY HO CHI MINH UNIVERSITY OF TECHNOLOGY FACULTY OF ELECTRICAL AND ELECTRONICS ENGINEERING DEPT OF CONTROL ENGINEERING AND AUTOMATION GRADUATION ESSAY DEVELOPMENT OF AUTOMATIC TEACHING METHOD USING STEREO CAMERA FOR SCARA ROBOTS SVTH: Ph畉m Ph動畛c D滴ng 1710875 GVHD: TS. Nguy畛n Hong Gi叩p 26/12/2022
- 4. 01 Introduction Reasons for choosing the topic Thesis objectives System overview 26/12/2022 4
- 5. 01 Introduction 1.1. Reasons for choosing the topic 26/12/2022 5
- 6. 1.1. Reasons for choosing the topic 26/12/2022 6 The teaching-less method for robots is being strongly developed in the world because of the advantages compared to the traditional teaching ones. The successful development of this topic will increase the diversity of robot applications by solving problems that cannot be solved by traditional teaching methods. Remote robot-assisted surgery Robot welding
- 7. 01 Introduction 1.2. Thesis objectives 26/12/2022 7
- 8. 1.2. Thesis objectives Develop a teaching-less method using stereo camera for SCARA robot to perform the pick-and-place feature with high precision. Build a camera - robot system which can operate based on designed algorithms and principles. 26/12/2022 8 Pick-and-place feature with high precision Force Torque Sensor required
- 9. 01 Introduction 1.3. System overview 26/12/2022 9
- 10. 1.3. System overview Principle of system operation 26/12/2022 10
- 11. 1.3. System overview Hardware system structure 26/12/2022 11
- 12. 1.3. System overview Principle of component operation 26/12/2022 12
- 13. Hardware System Design Design and build hardware system 02 26/12/2022 13
- 14. 02 Hardware System Design Marker Passive marker Stereo camera Stereo camera mounted from 2 Basler Dart daA1600-60um Module emits IR IR Filter H畛 robot SCARA Robot SCARA Epson EC251S DTP7H-coreCon robot controller 26/12/2022 14
- 15. 02 Hardware System Design 2.1. Marker 26/12/2022 15
- 16. 2.1. Marker 26/12/2022 16 Robot Marker Passive marker Setpoint Marker Robot Marker Setpoint Marker Length (mm) 60 98.2 Width (mm) 60 87 Height (mm) 75 22.7 Markers size
- 17. 02 Hardware System Design 2.2. Stereo camera 26/12/2022 17
- 18. 2.2. Stereo camera Stereo camera mounted from 2 Basler Dart daA1600-60um Resolution: 2x(640x480) pixel. Max Frame: 45FPS. Accuracy: 0.25mm RMS Operating distance: 500mm to 1500mm 26/12/2022 18
- 19. 02 Hardware System Design 2.3. Robot SCARA 26/12/2022 19
- 20. 2.3. Robot SCARA Robot SCARA EC251S DTP7H-coreCon robot controller 26/12/2022 20
- 21. Software Design Build algorithms and principles for model 03 26/12/2022 21
- 22. 03 Software Design Algorithm design Thread for position calculating Thread for end-effector's angle updating Thread for communication between PC and robot controller Convert distance User interface design Main program Error evaluation experimental program Communicate with PC using robot controller Setup communication protocol on robot controller 26/12/2022 22
- 23. 03 Software Design 3.1. Algorithm design 26/12/2022 23
- 24. 3.1. Algorithm design Thread for position calculating 26/12/2022 24 ヰ倹＄ = (ヰ＄ ヰ)2+ (＄ )2+ (ю＄ ю)2
- 25. 3.1. Algorithm design Thread for end-effector's angle updating 26/12/2022 25
- 26. 3.1. Algorithm design Thread for end-effector's angle updating 26/12/2022 26 Rotation matrix obtained from Homogeneous matrix ＄ : (, , ) = 11 12 13 21 22 23 31 23 33 So the pitch angle is: = arctan(31, 32 2 + 33 2 ) (rad) Calculation of rotation angle difference between two markers: ワ艶基 = 180 (艶諮 艶＄) (degree)
- 27. 3.1. Algorithm design Thread for communication between PC and robot controller 26/12/2022 27
- 28. 3.1. Algorithm design Thread for communication between PC and robot controller 26/12/2022 28 Time delay for robot moving to Setpoint position: ｐ = 倹＄ 100 ヰ ｐ 1000 + 500 - Robot maximum moving speed is set in robot controller to 1000 (mm/s). - Speed entered by the user, default will be 10 (%). - Since the robot has time to accelerate and decelerate in each movement and this value cannot be interfered with or measured, so in this project, I will use 500 (ms) to characterize this time value.
- 29. 3.1. Algorithm design Convert distance using equivalent coordinate system 26/12/2022 29 Incremental value xRobot = - Incremental value xCamera Incremental value yRobot = - Incremental value zCamera Incremental value zRobot = Incremental value yCamera
- 30. 03 Software Design 3.2. User interface design 26/12/2022 30
- 31. 3.2. User interface design Main program Module Tracker Module Pick Place App 26/12/2022 31
- 32. 3.2. User interface design Error evaluation experimental program 26/12/2022 32
- 33. 03 Software Design 3.3. Communicate with PC using robot controller 26/12/2022 33
- 34. 3.3. Communicate with PC using robot controller Config IP Address and Port on robot controller 26/12/2022 34 Programing on robot controller Config IP Address
- 35. Experiments and Evaluation Experimental results to verify the operability of model 04 26/12/2022 35
- 36. 04 Experiments and Evaluation 26/12/2022 36 Setup camera Setup robot and experimental platform
- 37. 04 Experiments and Evaluation 26/12/2022 37 Setup experimental environment
- 38. 04 Experiments and Evaluation Determining repeatability Checking distance correlation Comparing feedback position Evaluating the operability of model 26/12/2022 38
- 39. 04 Experiments and Evaluation 4.1. Determining repeatability 26/12/2022 39
- 40. 4.1. Determining repeatability Mica panel with holes cut by laser 26/12/2022 40 Marker fit perfectly inside mica panel
- 41. 4.1. Determining repeatability Markers position unchanged 26/12/2022 41 Capture markers position using software
- 42. 4.1. Determining repeatability Deviation from the standard position (mm) Repeatability (mm) Min Max 1 0.012 0.088 0.045 2 0.016 0.273 0.078 3 0.022 0.116 0.051 Experiment results 26/12/2022 42
- 43. 04 Experiments and Evaluation 4.2. Checking distance correlation 26/12/2022 43
- 44. 4.2. Checking distance correlation 26/12/2022 44 Fixed marker to robot end-effector Get end-effectors position from robot controller Get markers position from software
- 45. 4.2. Checking distance correlation Experiment results Distance difference between 2 points determined by robot controller (mm) Distance difference between 2 points determined by software (mm) Error (mm) Mean Error (mm) RMS Error (mm) 1 21.723 21.647 0.076 0.156 0.195 2 11.099 11.134 0.035 3 37.729 38.056 0.327 4 15.519 15.501 0.017 5 17.402 17.413 0.011 6 39.562 39.457 0.105 7 33.119 32.774 0.345 8 24.685 24.418 0.268 9 40.316 40.450 0.134 10 35.143 34.942 0.200 11 21.504 21.506 0.001 12 30.915 31.158 0.242 13 25.841 25.630 0.210 14 23.920 23.616 0.304 15 11.234 11.297 0.063 26/12/2022 45
- 46. 04 Experiments and Evaluation 4.3. Comparing feedback position 26/12/2022 46
- 47. 4.3. Comparing feedback position Compare Setpoints position sent and Robots position received 26/12/2022 47 Graphing and calculating errors
- 48. 4.3. Comparing feedback position Experiment results With 4 Setpoint unchanged position 26/12/2022 48
- 49. 4.3. Comparing feedback position Experiment results With 4 Setpoint unchanged position 26/12/2022 49
- 50. 4.3. Comparing feedback position Experiment results With 4 Setpoint unchanged position 26/12/2022 50
- 51. 4.3. Comparing feedback position Experiment results With 4 Setpoint unchanged position 26/12/2022 51
- 52. 04 Experiments and Evaluation 4.4. Evaluating the operability of model 26/12/2022 52
- 53. 4.4. Evaluating the operability of model Video to demo the operability of model 26/12/2022 53
- 54. Conclusion and Development direction Evaluate the completeness and feasibility of the model 05 26/12/2022 54
- 55. 05 Conclusion and Development direction Conclusion Development direction 26/12/2022 55
- 56. 05 Conclusion and Development direction 5.1. Conclusion 26/12/2022 56
- 57. 5.1. Conclusion Achievements Complete the algorithm to calculate the displacement distance and update the rotation angle of SCARA robot. Complete the system with Master and Slave stations to perform the pick-and-place feature with SCARA robot. Complete the communication protocol between PC and SCARA robot controller. Complete user interface. Complete experiments to verify the operability of model. 26/12/2022 57
- 58. 5.1. Conclusion Difficulties Due to limitation on the markers surface which can reflect IR (3 out of 4), in case SCARA robot rotate Robot Marker with the blind surface (the surface have no reflecting point) toward camera, sorfware will unable to calculate and update the end-effectors rotation angle. Havent optimized the synchronization solution between the robot's moving time and the cameras capture time to update the robot's rotation angle. 26/12/2022 58
- 59. 05 Conclusion and Development direction 5.2. Development direction 26/12/2022 59
- 60. 5.2. Development direction Optimize the synchronization solution between the robot's moving time and the cameras capture time. Replace the other marker with no blind surface. Upgrade the robot system (controller, driver, robot) to minimize the mechanical error. Build a response system from Slave to Master to inform the current status of robot (position, velocity, rotation,). Build a controller mounted directly to camera in order to reduce systems size. 26/12/2022 60
- 61. Thank you so much for your interest and attention!
- 62. Principle of marker recognition 26/12/2022 62 3D reconstruction of balls in DRB : Comparing geometry consists of balls in DRB -> Center position, distance between closer balls, etc. Rom file data of balls in DRB (DRB)