Contrastive Label Disambiguation for Self-Supervised Terrain Traversability Learning in Off-Road Environments

TL;DR

This paper introduces a self-supervised learning framework for terrain traversability in off-road environments, using contrastive label disambiguation to improve accuracy without human annotations.

Contribution

It proposes a novel contrastive label disambiguation mechanism that iteratively refines pseudo labels for self-supervised terrain traversability learning.

Findings

Effective on RELLIS-3D dataset

Successful application to Gobi Desert dataset

Improves traversability prediction accuracy

Abstract

Discriminating the traversability of terrains is a crucial task for autonomous driving in off-road environments. However, it is challenging due to the diverse, ambiguous, and platform-specific nature of off-road traversability. In this paper, we propose a novel self-supervised terrain traversability learning framework, utilizing a contrastive label disambiguation mechanism. Firstly, weakly labeled training samples with pseudo labels are automatically generated by projecting actual driving experiences onto the terrain models constructed in real time. Subsequently, a prototype-based contrastive representation learning method is designed to learn distinguishable embeddings, facilitating the self-supervised updating of those pseudo labels. As the iterative interaction between representation learning and pseudo label updating, the ambiguities in those pseudo labels are gradually eliminated, enabling the learning of platform-specific and task-specific traversability without any human-provided annotations. Experimental results on the RELLIS-3D dataset and our Gobi Desert driving dataset demonstrate the effectiveness of the proposed method.