雙足機器人之步態補償控制設計

Abstract

本論文之目的在於設計一雙足機器人及其運動控制系統，並探討其於直線步行中遇到不平物體障礙時之步態補償以避免傾倒。藉由安裝在腳底的壓力感測器取得腳底受力的資訊，使雙足機器人得以在地面上有不平物體之步行中，產生適度的補償動作以完成步行。本論文中採用靜態步行(Static Walking)的控制方式來設計機器人的動作與軌跡，使用不同的參數設定可對軌跡進行修改。在方法上，本論文發展出一套補償控制器演算法，使用量測式的ZMP計算，於步行中可以藉由機器人腳底壓力感測器所獲得的ZMP資訊，判斷腳底障礙物的存在與位置，並因此產生機器人腳部的補償動作以進行障礙物之適應，使機器人得以順利跨越障礙物。文中以實驗室自行設計之60公分高的12自由度雙足機器人搭配工業電腦，驗證所提出方法的可行性。經過實驗驗證，此系統可相當程度提高機器人經過21mm高障礙物之穩定性。

The objective of this thesis is to design a biped robot and a motion controller such that it can handle an uneven floor in bipedal walking. Walking pattern for stably straight motion is generated online in real time. Using the information of the pressure sensors under the biped feet, the robot can walk across an uneven obstacle with suitable compensation of the angle of the ankle joint. The method of static walking is adopted to generate the walking trajectory which can be changed by modulating parameters. By measuring the ZMP, we also develop an algorithm which is used to sense obstacles under the foot and generating the compensation of the ankle. The compensation is combined with the walking trajectory to adapt to the obstacle on the ground while robot walking. Experimental results show that the biped robot can walk successfully across the obstacle which is lower than 21mm.