本研究介紹我們開發的一個適用於雙足人形機器人之拖曳式動作編修環境，以及我們如何改善拖曳時肇因於物理引擎的模擬穩定度。傳統的機器人動作編輯軟體只使用參數介面調整或捲軸控制器來調整每個關節的轉動角度，這個方式在動作製作上需花費相當龐大的時間。本研究開發的動作編修系統為直覺化與機器人之互動編修動作環境，透過滑鼠於動作編輯器上之模型身體部位拖曳，即可改變連接的身體部位姿勢，還可以透過物理引擎播放動作的模擬結果，大幅增快動作製作之效率。

本系統利用知名的開放源碼物理引擎－Bullet Physics實現物理環境，在動作編輯環境可以觀察機器人自體碰撞之狀況，以便使用者調整動作來防範自體碰撞。我們使用了Kondo KHR-2HV雙足人形機器人來讀取系統產生之動作檔，並與系統的模擬情況做對照。由於物理引擎使用的計算方法在運動計算上會產生誤差，而這使得我們在編修環境中使用滑鼠拉動身體部位時會有拉長及抖動之問題。針對這些問題我們使用既有的post-stabilization方法做改善，在我們的實驗中平均抖動幅度約改善了75%，最大拉長量約減少83%。另外為了效率上的考量我們另外提出使用運動學的方式實作拉動手臂，可以完全避免手臂拉長及抖動的問題。

This master’s thesis introduces a motion editing environment for biped humanoid robots we have developed in order to realize interactive drag-and-drop motion editing. Simulation stability issues caused by dynamics engine are also studied and compensated. Traditional motion editors use one degree-of-freedom controls like scroll bars to adjust the rotation degrees of joints and this inefficient method makes motion edition a time-consuming process. The developed motion editing environment supports drag-and-drop feature by which users edit static poses of any assigned moment by dragging hand or other part of the robot body, shown as a 3D model, and the whole limb moves according to inverse kinematics calculation. The inputted motions can then be verified immediately by motion playback. The developed motion editor is believed to significantly enhance the efficiency of motion editing.

We used the famous open source project Bullet physics engine to realize the physical simulator. It supports collision detection feature which allows users to modify motions in order to avoid self-collision. The Kondo KHR-2HV biped humanoid robot was used to test the edited motions. The algorithm Bullet uses to calculate motions of rigid body with kinematic constraints give results containing numeric errors in some dragging operations. These errors detach the dragging body parts or make part of the link continuously trembling. We used Post-stabilization method to reduce about 75% of trembling degrees and 83% of body gap errors. Furthermore, we propose a kinematic motion calculation method which can complete eliminate the above-mentioned problems in some cases.

[1] Yoshizaki, W., Sugiura, Y., Chiou, A., C., Hashimoto, S., and Inami, M., “An Actuated Physical Puppet as an Input Device for Controlling a Digital Manikin”, Proceedings of the ACM SIGCHI conference on Human factors in computing systems, New York, USA, pp. 637-646 (2011)
[2] Michel, O., “WebotsTM: Professional Mobile Robot Simulation”, Journal of Advanced Robotic Systems, Vol. 1, No. 1, pp. 39-42 (2004)
[3] Jackson, J., “Microsoft Robotic Studio: A Technical Introduction”, IEEE Robotics and Automation Magazine, Vol. 14, No. 4, pp. 82-87 (2007)
[4] Freese, M., Singh, S., Ozaki, F., and Matsuhira, N., “Virtual Robot Experimentation Platform V-REP: a Versatile 3D Robot Simulator”, Proceedings of the Second international conference on Simulation, modeling, and programming for autonomous robots, pp. 51-62 (2010)
[5] Koenig, K., Howard, A., “Design and Use Paradigms for Gazebo, An Open-Source Multi-Robot Simulator”, Proceedings of IEEE Intelligent Robots and Systems international Conference, pp. 2149-2154 (2004)
[6] Lewis, M., Wang, J., Hughes, S., “USARSim: Simulation for the Study of Human-Robot Interaction”, Journal of Cognitive Engineering and Decision Making, Vol. 1, No. 1, pp. 98-120 (2007)
[7] Diankov, R., Kuffner, J., “OpenRAVE: A Planning Architecture for Autonomous Robotics”, Technical report, Robotics Institute, Pittsburgh, PA (2008)
[8] Boeing, A., “Design of a Physics Abstraction Layer for Improving the Validity of Evolved Robot Control Simulations”, Ph.D. dissertation, The University of Western Australia, Crawley, Western Australia (2009)
[9] Erleben, K., “Stable, Robust, and Versatile Multibody Dynamics Animation”, Ph.D. dissertation, University of Copenhagen, Denmark (2005)
[10] Gilbert, E. G., Johnson, D. W., and Keerthi, S. S., “A Fast Procedure for Computing the Distance Between Complex Objects in Three-Dimensional Space”, IEEE Journal of Robotics and Automation, Vol. 4, No. 2, pp. 193-203 (1988)
[11] Van Den Bergen, G., “Proximity Queries and Penetration Depth Computation in 3D Game Objects”, Proceedings of Game Developers Conference, (2001)
[12] Mirtich, B. V., “Impulse Based Dynamic Simulation of Rigid Body Systems” , Ph.D. dissertation, University of California, Berkeley, USA (1996)
[13] Cottle, R. W., Pang, J. S., and Stone, R. E., The Linear Complementarity Problem, Society for Industrial and Applied Mathematics, Philadelphia, USA (2009)
[14] Teng, W. C., Wang, C. H., Yang, K.H., and Wang, P.J., “Development of an Interactive Dynamic Motion Editing Toolkit for Biped Humanoid Robots” , Proceedings of the 43th International Symposium on Robotics , Taipei, Taiwan (2012)
[15] J. Baumgarte, “Stabilization of Constraints and Integrals of Motion in Dynamical Systems”, Journal of Computer Methods in Applied Mechanics and Engineering, Vol. 1, pp. 1-16 (1972)
[16] Chang, C. O., and Nikravesh P.E., “An Adaptive Constraint Violation Stabilization Method for Dynamic Analysis of Mechanical Systems”, Journal of Mechanisms, Transmissions, and Automation in Design, Vol. 107, pp. 488–492 (1985)
[17] Cline, M. B., and Pai, D. K., “Post-Stabilization for Rigid Body Simulation with Contact and Constraints”, In Proceeding IEEE International Conference on Robotics and Automation, Taipei, Taiwan, Vol. 3, pp. 3744-3751 (2003)