A humanoid robot localization method for biped navigation in human-living environments

In real-time footstep planning for the humanoid robot, it is needed to obtain the accurate location of its foot placements. However, existing methods only needed to obtain the localization of the robot as a particle, which did not specifically consider the localization of its foot placements. This paper presents foot placement localization method for humanoid robot in human-living environments. A Kinect sensor was first used to obtain the digital map of the global environment and the position and orientation of the head of the robot was then extracted. The position and orientation of the foot placement was then calculated based on the spatial geometric relationship between the head and the target foot. The proposed method was used in the real time footstep planning in dynamic human-living environments. Experimental results validated its feasibility.