Dense RGB-D SLAM for Humanoid Robots in the Dynamic Humans Environment

Abstract

These two problems block the SLAM method applications for humanoids. On the one hand, humans are often considered as moving obstacles or moving targets in the humanoids working spaces, which result the dynamic environment problem. On the other hand, the disturbances caused by the executions of biped locomotion and the environment structure discontinuity caused by the falling down case make big challenge for SLAM approaches. In this paper, we propose a robust dense RGB-D environment reconstruction method for humanoids working in dynamic humans space. The proposed approach efficiently detects humans and fast reconstructs the static environments through deep learning-based human body detection, and then implement a graph-based segmentation on the RGB-D point clouds, which separates detected moving humans from the static environment. Finally, the separated static environments are aligned with using state-of-the-art frame-to-model scheme. Experimental results on both public benchmark and a newly developed HRP-4 humanoids SLAM dataset indicate that the proposed approach achieves outstanding performance in full dynamic environments.