在本篇論文裡, 我們使用模糊推論來動態選擇PID 控制器之參數, 控制一個半球機器人使其能達到自我平衡控制目的。為了能使上方自由擺動以保持平衡, 我們將機器人下方設計成半球狀, 以得到兩個自由度(左右擺動及前後擺動), 為了能夠監控半球機器人的傾斜角度, 我們在馬達底部安裝一個陀螺儀, 用來監控。並設計了二個馬達來分別控制二個方向的擺動。一般我們做平衡控制, 都是維持上方不動, 下方來移動做調整, 本篇論文裡使用相反的設計, 下方會隨著上方做擺動而跟著擺動。經由Recursive NEWTON-EULER Formulation [1], 我們推導出半球機器人的動態方程式, 此方程式可用來模擬半球機器人的運動狀態, 若再將迴授誤差訊號輸入至控制器, 即可使閉迴路控制系統保持穩定並保持平衡。最後, 我們將使用MATLAB 來模擬半球機器人的平衡控制以證明所提出的控制方法為有效且正確的。
In this thesis, we establish a hemisphere robot to achieve automatic balance control, and use PID controller with FUZZY inference to control the robot. In general, to maintain balance vertically for the robot and does not to move give a challenge in control. The hemisphere robot uses a balance beam on the top of the end effector taking right and left swings by keeping the beam on the horizontal axis. In order to swing freely to achieve balance, two AI motors are setup for the hemisphere robot to have two degrees of freedom ( swinging back and forth ), and a gyroscope with x−y −z tilt angles is installed at the bottom of the robot. Then, x−y−z tilt angles are fedback through inverse kinematics to the controller to actuate two motors such that the beam is kept on the horizontal axis after swinging. By using the Recursive NEWTON-EULER Formulation [1], we derive the hemisphere robot dynamic equations used to simulate the dynamic motion of the hemisphere robot. Finally, some simulations for the hemisphere robot using MATLAB is used to demonstrate the effectiveness and correctness of the proposed control method.