| 摘要: | 本論文目標在於設計及製作出一台具有強健平衡功能的雙足機器人。首先，我們製作出第一代機器人，並利用機器人多連桿系統與COP 平衡之概念，設計出「姿態平衡控制器」，藉此使用者只需輸入機器人的「初始」及「目標」兩筆動作矩陣至控制器內，控制器就會透過內插法、COP 穩定判斷與腳踝修正機制，將不穩定的動作矩陣調整為穩定動作軌跡，減輕使用者設計動作的負擔。第一代機器人能平穩的執行舉手、踢腳、鞠躬、蹲下及行走等動作。接著，我們將第一代機器人做硬體及控制器的改良，完成了第二代機器人。硬體架構上增加多維度腰部機構，讓機器人能以更多元之方式調整姿態、適應環境，接著運用空間感測器偵測地面傾斜程度，並搭配模糊控制器的即時運算，修正機器人在斜坡上的不穩定姿態，再以影像處理技術進行目標物辨識、追蹤與射擊。基於以上，第二代機器人能夠穩定執行舉重、抓取單槓且平移、斜坡行走平衡與瞄準射擊氣球等任務，充分表現出機器人在平衡上的強健性，與影像處理的高辨識率。最後由實驗數據分析證明，腰部機構與模糊控制器對於機器人平衡姿態，具有極佳的貢獻。The objective of this study is to design and fabricate a robust balance biped robot. First, the “first generation humanoid biped robot” is constructed, and then the “balance controller” is designed by using the multi-link system and the COP concept. The users just input the initial state and final state motion matrices into the controller, based on the linear interpolation, COP concept, and ankle modification method, the motion matrices can be adjusted to a stable motions trajectory such that the humanoid robot can do the motion trajectory and keep balance all the time. The first generation humanoid robot can stably perform the motions such as raising hand, kicking, bowing, squatting, and walking. Subsequently, the second generation humanoid robot is accomplished by improving the hardware and controller of the first generation humanoid robot. Three degrees of freedom (DOF) motors are set at the waist, inertial measurement unit (IMU) and the fuzzy controller are employed to make the robot adjust its posture and then stably stand on the slope platform. Furthermore, this robot has image processing function to recognize, track, and shoot the target. In summary the second generation humanoid robot can accomplish the motions such as weightlifting, grabing horizontal bar, walking on the slope platform and shooting target, which show the effectiveness of robust balance control and image recognition. Finally, the experimental result data show that the waist structure and fuzzy controller can help the humanoid robot to keep balance obviously. |
