Abstract: Aiming at the problems of NAO robot walking slowly, walking deviation course caused by motor error, fever and other factors in the maze, a fast inflection point and forward heading correction algorithm is proposed. Firstly, rapid turning point detection is performed in complex indoor environments. Then the geometric model is constructed according to the imaging principle of the camera. Based on this model, the distance between the robot and the inflection point is deduced from the geometrical relation of the image, and then the robot can reach the inflection point quickly by rotating a certain angle according to the included angle. Experiments show that the proposed method can accurately prevent the deviation of NAO, and the deviation rate is less than 0.4%. At the same time, NAO’s walking rate is increased by about 34.8%.

Key words: fast inflection point distance detection, NAO robots, correcting the deviation, basis of image geometric relation, inflection point detection

摘要： 针对NAO机器人在迷宫中行走缓慢以及因电机误差、发热等因素导致的行走偏离航线等问题，提出一种快速拐点测距和前进航向矫正算法。首先，在室内复杂环境下进行快速拐点检测。接着按照摄像机成像原理构建几何模型，在此模型上进行图像几何关系推导,得到目前机器人距离拐点的距离，进而根据偏离夹角旋转一定的角度，控制机器人快速到达拐点。实验证明，所提出的方法可以准确防止机器人走偏，偏差率低于0.4%，同时NAO的行走速率提高了大约34.8%。

关键词: 快速拐点测距, NAO机器人, 航向矫正, 图像几何关系, 拐点检测