Measurement of Ground Surface for Walking on Rough Terrain with Wearable Robot

In this paper, we propose a floor surface measurement system for walking assistance with a wearable robot. The point cloud of the floor, user's feet, and an obstacle were measured by a RGB-Depth camera. We assume that most of the points in the point cloud consist of the surface of the floor, and the plane of the floor surface was reconstructed by a principal component analysis (PCA) of the point cloud. The obstacle was reconstructed by the subgroup of the points cloud in which the distance from the floor surface was higher than a threshold. The experimental results showed that the PCA-based method could discriminate the floor surface and obstacle and could estimate the size and location of the obstacle.