臺灣博碩士論文加值系統

本論文主要針對工業機械手臂之機構參數誤差進行校準與補償，量測設備使用LEAP MOTION LM-010感測器，其優點為價格便宜、體積小、操作方便、開放程式碼等，以及可接受範圍之量測精度(0.01mm)；因此，本論文之重點在於評估一低成本、非接觸式以及可量產之應用在工業機械手臂的定位精度校正量測設備的可行性。

本研究目標找出實際的機構參數，並利用加工程式碼進行補償，在手臂校準的等級上歸類為第二級校準(2nd-level calibration)。為了要校準機械手臂，本研究使用Denavit–Hartenberg(D-H) method建立順向運動模型(Forward kinematic model)、幾何近似法(Geometric approach)建立反向運動模型，以LEAP MOTION sensor作為量測設備取得量測資料，並以最小平方誤差法計算機構參數誤差量，以及使用Fake pose method進行旋轉軸之旋轉角度補償，最後以ISO9283:1998為標準，驗證手臂校準前與校準後的姿態精度(Pose accuracy, AP)。

本論文主要針對工業機械手臂之機構參數誤差進行校準與補償，量測設備使用LEAP MOTION LM-010感測器，其優點為價格便宜、體積小、操作方便、開放程式碼等，以及可接受範圍之量測精度(0.01mm)；因此，本論文之重點在於評估一低成本、非接觸式以及可量產之應用在工業機械手臂的定位精度校正量測設備的可行性。

本研究目標找出實際的機構參數，並利用加工程式碼進行補償，在手臂校準的等級上歸類為第二級校準(2nd-level calibration)。為了要校準機械手臂，本研究使用Modified Denavit–Hartenberg(D-H) method建立順向運動模型(Forward kinematic model)、幾何近似法(Geometric approach)建立反向運動模型，以LEAP MOTION sensor作為量測設備取得量測資料，並以最小平方誤差法計算機構參數誤差量，以及使用Fake pose method進行旋轉軸之旋轉角度補償，最後以ISO9283:1998為標準，驗證手臂校準前與校準後的姿態精度(Pose accuracy, AP)。

在校準前，以ISO test確認機械手臂之AP，量測範圍為90立方毫米之立方體，量測點5個，重複30次循環，得到AP=0.34mm之結果；使用數學軟體MATLAB進行誤差模型之模擬驗證，結果證實可以最小平方誤差法得到機構參數之誤差，以及使用fake pose method進行旋轉角補償角度之計算；惟經過實驗確認，LEAP MOTION雖然官方量測精度達0.01毫米，但距離Y方向越遠量測精度越差，線性度方面，X、Z方向線性度最大變動値約1mm、Y方向線性度最大變動値約2.5mm。將Leap motion量測的機器手臂空間位置，進行機器手臂參數估測，發現機構參數迭代計算擬合雖然收斂，但是得到不合理的機構參數結果，推測有三，一為可能是LEAP MOTION量測變動値過大、二為儀器架設誤差造成，或是機構參數在數值估算時收斂到不合理的解。

This thesis is focusing on kinematic parameter calibration of industrial robot using LEAP MOTION sensor (model number LM-010) as measurement device. The sensor has several advantages, such as low cost, small, easy to use and open code, etc. In short, the point of this thesis is to evaluate a way of low cost, and contactless sensor for calibrating industrial robot kinematic parameter.

The level of this calibration is defined 2nd –level calibration. To calibrate robot, Denavit-Hartenberg (D-H) method has been used as forward kinematic model, Geometric approach as inverse kinematic model and LEAP MOTION sensor as measuring device. After measuring, least square method has been used to identified errors in kinematic parameter and fake pose method to calibrate in angle of revolute joints. In the end, ISO9283:1998 is used to check the result after calibration and compensation.

Before calibration, ISO9283:1998 has been used to check pose accuracy of the robot in five measuring point and thirty cycles in 60mm3 cube, and the result is 0.34mm. Kinematic parameters have been simulated and estimated in MATLAB software. The simulation result has proven the feasibility of this proposed method that using least square method to estimated kinematic parameters and fake pose method to compensate in angles of revolute joints. Although nominal measuring repeatability of the leap motion sensor is claimed to be 0.01mm, but it has been found that the experimental results have much worse value especially when measuring position in the Y direction in farther positions. In linearity, X and Z direction has maximum deviation about 1.00mm, and 2.50mm in Y direction. The proposed least square method has failed to estimate robot kinematic parameters because the computational results is not reasonable. The reasons may be from the setting error of the measuring device or from the not enough repeatability of the leap motion sensors or getting unreasonable solution even it was convergence.

摘要 I
ABSTRACT III
目錄 V
圖目錄 VII
表目錄 XIII
一、 緒論 1
1.1. 研究動機 1
1.2. 論文架構 1
1.3. 研究目標 2
1.4. 文獻回顧 2
1.5. 專利分析 7
1.6. 與前人的傳承關係創新之處 16
二、 原理簡介 17
2.1 六軸機械手臂順向運動模型 17
2.1.1 Denavit–Hartenberg Method 17
2.1.2 Euler Convention 20
2.1.3 Yaw, Pitch, Roll Convention 23
2.1.4 Staubli TX90 順向運動模型 25
2.2 六軸機械手臂反向運動模型 28
2.2.1 Kinematic Decoupling 28
2.2.2 Inverse Position : Geometric Approach 29
2.2.3 Inverse Orientation 33
2.2.4 Staubli TX90 反向運動模型 34
2.3 最小平方法 37
2.4 FAKE POSE COMPENSATION 40
2.5 機械手臂校準原理 41
2.6 模型模擬與驗證 42
2.6.1 正向運動模型 42
2.6.2 反向運動模型 43
2.6.3 機構參數誤差模型 47
2.7 LEAP MOTION LM-010量測原理 64
2.8 ISO 9283:1998 65
2.8.1 量測路徑 65
2.8.2 計算方式 66
三、 實驗架構與儀器介紹 68
3.1 量測儀器設備 68
3.1.1 Staubli TX90 Robot 68
3.1.2 LEAP MOTION LM-010 Sensor 73
3.1.3 End-effector 74
3.2實驗流程 76
四、 實驗結果與討論 77
4.1 LEAP MOTION LM-010量測精度實驗 77
4.2 ISO TEST實驗結果 99
4.2.1 校準前量測結果 100
4.3 校準路徑量測結果 104
4.4 最小平方法擬合結果 118
五、結論與未來展望 122
5.1 研究結果與討論 122
5.2 未來展望 123
參考文獻 124

Uncategorized References
[1]李勇志, "三維雷射量測儀之研發及在工業機器人校正之應用," 碩士, 機械工程學研究所, 國立臺灣大學, 台北市, 1999.
[2]J. P. Prenninger, "Contactless position and orientation measurement of robot end-effectors," in Robotics and Automation, 1993. Proceedings., 1993 IEEE International Conference on, 1993, pp. 180-185.
[3]A. Goswami, A. Quaid, and M. Peshkin, "Identifying robot parameters using partial pose information," Control Systems, IEEE, vol. 13, pp. 6-14, 1993.
[4]J. M. S. T. Motta and R. S. McMaster, "Experimental validation of a 3-D vision-based measurement system applied to robot calibration," Journal of the Brazilian Society of Mechanical Sciences, vol. 24, pp. 220-225, 2002.
[5]陳建勳, "使用影像量測系統校準模組化機器人," 碩士, 機械工程研究所, 大同大學, 台北市, 2003.
[6]T. Kivelä, H. Saarinen, J. Mattila, V. Hämäläinen, M. Siuko, and L. Semeraro, "Calibration and compensation of deflections and compliances in remote handling equipment configurations," Fusion Engineering and Design, vol. 86, pp. 2043-2046, 2011.
[7]A. Nubiola, "Calibration of a serial robot using a laser tracker," École de technologie supérieure, 2011.
[8]F. Weichert, D. Bachmann, B. Rudak, and D. Fisseler, "Analysis of the accuracy and robustness of the leap motion controller," Sensors, vol. 13, pp. 6380-6393, 2013.
[9]許哲勝, "多重智慧控制器應用於機械手臂定位," 碩士, 機電科技學系, 國立臺灣師範大學, 台北市, 2013.
[10]M. W. Spong, S. Hutchinson, and M. Vidyasagar, Robot modeling and control vol. 3: Wiley New York, 2006.
[11]R. Bernhardt and S. Albright, Robot calibration: Springer Science & Business Media, 1993.
[12]W. K. Veitschegger and C.-H. Wu, "Robot accuracy analysis based on kinematics," Robotics and Automation, IEEE Journal of, vol. 2, pp. 171-179, 1986.
[13]W. Veitschegger and C. H. Wu, "A method for calibrating and compensating robot kinematic errors," in Robotics and Automation. Proceedings. 1987 IEEE International Conference on, 1987, pp. 39-44.
[14]N. Silberman and R. Fergus, "Indoor scene segmentation using a structured light sensor," in Computer Vision Workshops (ICCV Workshops), 2011 IEEE International Conference on, 2011, pp. 601-608.
[15]A. Kolb, E. Barth, R. Koch, and R. Larsen, "Time-of-flight sensors in computer graphics," in Proc. Eurographics (State-of-the-Art Report), 2009.
[16]K. Ambrosch and W. Kubinger, "Accurate hardware-based stereo vision," Computer Vision and Image Understanding, vol. 114, pp. 1303-1316, 2010.
[17]E. ISO, "9283: 1998," Manipulating industrial robots-Performance criteria and related test methods, 1998.