Self-Supervised Traversability Prediction by Learning to Reconstruct Safe Terrain

作     者：Schmid, Robin Atha, Deegan Schöller, Frederik Dey, Sharmita Fakoorian, Seyed Otsu, Kyohei Ridge, Barry Bjelonic, Marko Wellhausen, Lorenz Hutter, Marco Agha-Mohammadi, Ali-Akbar 

作者机构： PasadenaCA United States Swiss Federal Institute of Technology ETH Zürich Robotic Systems Lab Switzerland University of Goettingen Department of Computer Science Germany Technical University of Denmark Department of Electrical Engineering and Photonics Denmark 

出 版 物：《arXiv》 (arXiv)

年 卷 期：2022年

核心收录：

主　　题：Off road vehicles 

摘      要：Navigating off-road with a fast autonomous vehicle depends on a robust perception system that differentiates traversable from non-traversable terrain. Typically, this depends on a semantic understanding which is based on supervised learning from images annotated by a human expert. This requires a significant investment in human time, assumes correct expert classification, and small details can lead to misclassification. To address these challenges, we propose a method for predicting high- and low-risk terrains from only past vehicle experience in a self-supervised fashion. First, we develop a tool that projects the vehicle trajectory into the front camera image. Second, occlusions in the 3D representation of the terrain are filtered out. Third, an autoencoder trained on masked vehicle trajectory regions identifies low- and high-risk terrains based on the reconstruction error. We evaluated our approach with two models and different bottleneck sizes with two different training and testing sites with a four-wheeled off-road vehicle. Comparison with two independent test sets of semantic labels from similar terrain as training sites demonstrates the ability to separate the ground as low-risk and the vegetation as high-risk with 81.1% and 85.1% accuracy. Copyright © 2022, The Authors. All rights reserved.