Study of Supervised Learning based Traversabilty Perception for Self-Supervised Learning of Perceptual System

Abstract

In previous works, we have proposed a traversability perception method based on Growing Neural Gas with Different Topologies for the space perception of autonomous mobile robots. However, the proposed method uses pre-designed parameters such as slope angle and surface roughness. In this study, we propose a method for estimating traversability based on supervised learning as a method for learning traversability perception.