- 學位論文
雙足機器人之步態規劃與感測系統建置
Gait generation of a biped robot and construction of measurement system
摘要
本文主要的目的是要設計出一個能使機器人穩定行走的步行模式，並使用新設計的九軸慣性量測系統量測機器人步行時的質心線加速度和角加速度。 在機器人的運動學上，我們定義了用來描述機器人每個關節之間的參數，並確立了馬達轉角方向和地座標之間的關係。也推導了能夠轉換機器人質心運動空間（Working Space）和馬達轉角運動空間（Joint Space）的雙向轉換方程式。在機器人的動力學上，引進了零矩合點（Zero-Moment Point, ZMP）的概念，確立了步行時能使機器人穩定的條件。我們將機器人簡化成倒單擺，加上適當的線性化和簡化假設，用力學平衡推導出步行時機器人的質心軌跡方程式，也推導出在質心軌跡和ZMP軌跡之間轉換的方法。在機器人步行上，將一個完整的步行區分成數個步行階段，每個階段依據其階段特性用不同的動力學方程式推導出質心軌跡和腳掌軌跡，並透過逆運動學求出馬達轉角軌跡。最後求出在此馬達轉角軌跡下的ZMP軌跡，並確認是否符合平衡條件。有了馬達轉角軌跡後，先用Matlab做運動學上的模擬，確定馬達轉角軌跡的正確性後，再用Adams做動力學上的模擬，觀察加入重力影響後是否還能平穩的步行，最後再透過實驗來驗證整個步行軌跡產生的有效性。 在量測機器人動態資訊上，我們開發出了九軸慣性量測系統，透過安裝在非質心位置上的三個二軸加速規和一個三軸陀螺儀，可以推算出質心的線加速度和雜訊較傳統方法小的角加速度，最後用一個二軸傾斜儀消除重力對加速規量測值的影響。
並列摘要
This thesis discusses the design and realization of stable walking pattern for a humanoid biped robot, and the development of 9-axis IMU (inertial measurement unit) system which is able to derive linear and angular accelerations of COG (Center Of Mass) of the robot. In the kinematics of robot, we defined the parameter to describe to relationship each links on the robot, and derived the relationship between rotational direction of each motor and the ground coordinate. Next, we deal with the transformation equations between the Joint Space and the Working Space. In the dynamics of robot, we introduce the concept of ZMP (Zero-Moment Point), and analysis the condition for a stable walking of humanoid robot. We simplified the robot to an inverted pendulum model, and derived the COG trajectory base on reasonable assumptions. We also derived the transformation equations between the COG trajectory and the ZMP trajectory.We separated a full walking into several stages, including squat down stage and stand up stage. According to characteristics of each stage, different dynamic model is applied to derive the COG and feet trajectories. Next, the angular positions of motors are derived through inverse kinematics. Furthermore, the ZMP trajectory was derived by the motor trajectory through forward kinematics, and was proved to satisfy the condition of stable walking.We first simulated the motor trajectory by Matlab, and approve it is reasonable in kinematics. Then we simulated the motor trajectory by Adams to approved it generate a stable walking of robot in dynamics. After simulations, the trajectory was applied to real robot and to walk in real environment. To obtain the information of robot, we develop the 9-axis IMU (Inertial Measurement Unit) system. Three 2-axis accelerometers and one 3-axis rate gyro are installed to measure 3-axis linear accelerations and 3-axis angular accelerations of COG. The 9-axis IMU system releases the limitation that the tradition IMU is required to install on COG, and the 9-axis IMU system obtain better quality of signal than the tradition IMU. Besides, a 2-axis inclinometer is applied to compensate the effect of gravity.