人型機器人之穩定動作設計與控制

Stable Motions Design and Control in Humanoid Robot

劉欣宇(Hsin-yu Liu)

指導教授 ： 王文俊

國立中央大學/資訊電機學院/電機工程學系/博士(2012年)

摘要

本論文提出了人型機器人的穩定動作設計與動態平衡控制，並應用於實際的機器人上以驗證提出方法之效能。在人型機器人的穩定動作設計方面，首先以模擬的方式建立人型機器人的多連桿系統，接著隨時計算機器人在動作時的零力距點(zero moment point)位置，並修正機器人腳踝的角度以確保其在連續動作時其零力距點皆落於穩定範圍內，最後將調整過的機器人動作套用於實際人型機器人上以驗證其穩定性。在人型機器人的動態平衡控制方面分為兩個階段，首先以設置於機器人腳底的壓力感測器為輸入，以腳踝模糊控制器調整機器人的腳踝關節，再次確保機器人在動作時的穩定。接下來配合感測上半身傾斜角度，再以腳踝-腰部模糊控制器同時控制機器人的腳踝與腰部關節以增加其抵抗外力的能力。最後提出三種實驗，讓機器人抵抗三種不同類型的外力，分別為抵抗槌子的推擠、站立於雙軸轉動的平台以及在蹺蹺板上行走，以測試控制器施於機器人的強健穩定性。而實驗數據顯示，所提出的模糊制器確實能夠大幅提升機器人對於外力抵抗的能力。總而言之，此論文對於小型人型機器人提供了一套完整的動作穩定方法，主要以模擬的方式先調整機器人的動作，並利用感測器回授及控制器設計，使機器人能夠即時的達成平衡控制，進而增加其抵抗外力的能力。最後再以實驗數據呈現出控制器所產生的效能。

關鍵字

動作設計 ； 人型機器人 ； 模糊控制

並列摘要

In this dissertation, the methods of stable motion design and balance control for a humanoid robot are proposed and the effectiveness of the methods is demonstrated by a practical humanoid robot. In the stable motion design, the concepts of multilink system and ZMP (Zero Moment Point) are applied to construct a virtual model of the humanoid robot for checking the ZMP positions of the robot’s motions. Subsequently, the stability of the virtual motions is checked by adjusting the angle of robot’s ankle and those stable motions are realized by a practical humanoid robot. In the balance control of the humanoid robot, the proposed method includes real-time balance control for the humanoid robot to imitate human motions and resist external uncertain factors. Regarding the real-time balance control for imitating human motions, the force sensors are installed under the soles of the robot and the fuzzy controller is applied to adjust the robot’s ankle joint in real time for increasing the stability of the motions. Regarding the real-time balance control for resisting external uncertain factors, the force sensor module and IMU (inertial measurement unit) sensor are used as the stability indexes and they are the inputs of four fuzzy controllers. By applying the fuzz controllers, the waist and ankle joints of the humanoid robot are adjusted simultaneously for resisting environmental disturbances. In the experiment, the humanoid robot is required to achieve three tasks which are resisting a thrust force from a hammer, standing on a two-axis inclinable platform and walking on a seesaw. The experimental data demonstrates that the proposed fuzzy controllers can indeed increase the stability of the humanoid robot. In summary, this dissertation proposes serial methods for stabilizing the motions of the humanoid robot. The robot’s motions are stabilized by the simulation method and then performed by a practical humanoid robot. Based on the sensor feedbacks and the proposed fuzzy controllers, the stability of the robot is checked again during the motions implementation. The proposed methods are successfully verified by the practical experiments.

並列關鍵字

motion design ； fuzzy control ； humanoid robot

參考文獻

[5] K. Erbatur, A. Okazaki, K. Obiya, T. Takahashi, and A. Kawamura, “A study on the zero moment point measurement for biped walking robots,” in proceeding of 7th international Workshop on Advanced Motion Control, Maribor, Slovenia, 2002, pp. 431-436.

[8] J. Or, “A hybrid cpg–zmp controller for the real-time balance of a simulated flexible spine humanoid robot,” IEEE Transactions on Systems, Man and, Cybernetics Part C-Application, vol. 39, no.5, pp. 547-560, Sep. 2009.

[9] J.P. Ferreira, M. M. Crisostomo, A. P. Coimbra, “SVR versus neural-fuzzy network controllers for the sagittal balance of a biped robot,” IEEE Transactions on Neural Networks, vol. 20, no. 12, pp. 1885-1897, Dec. 2009.

[10] F. Kanehiro, M. Inaba, and H. Inoue, “Development of a two-armed bipedal robot that can walk and carry objects,” in proceeding of International Conference on Intelligent Robots and Systems, Osaka, Japan, 1996, vol.1, pp. 23–28.

[12] J. G. Juang, “Fuzzy neural network approaches for robotic gait synthesis,” IEEE Transactions on Systems, Ma and, Cybernetics Part B- Cybernetics, vol. 30, no. 4, pp. 594-601, Aug. 2000.