Humanoid Robot Navigation: Getting Localization Information from Vision

In this article, we present our work to provide a navigation and localization system on a constrained humanoid platform, the NAO robot, without modifying the robot sensors. First we try to implement a simple and light version of classical monocular Simultaneous Localization and Mapping (SLAM) algorithms, while adapting to the CPU and camera quality, which turns out to be insufficient on the platform for the moment. From our work on keypoints tracking, we identify that some keypoints can be still accurately tracked at little cost, and use them to build a visual compass. This compass is then used to correct the robot walk, because it makes it possible to control the robot orientation accurately.