Learning Off-Road Terrain Traversability with Self-Supervisions Only

TL;DR

This paper presents a self-supervised approach for off-road terrain traversability estimation that learns from past driving trajectories without manual labels, enabling reliable predictions across diverse environments.

Contribution

It introduces a novel self-supervised learning framework combining trajectory-based labels and visual representation learning for terrain traversability estimation.

Findings

Outperforms other self-supervised methods in various environments

Achieves comparable performance to supervised methods with manual labels

Demonstrates reliable traversability estimation in diverse off-road conditions

Abstract

Estimating the traversability of terrain should be reliable and accurate in diverse conditions for autonomous driving in off-road environments. However, learning-based approaches often yield unreliable results when confronted with unfamiliar contexts, and it is challenging to obtain manual annotations frequently for new circumstances. In this paper, we introduce a method for learning traversability from images that utilizes only self-supervision and no manual labels, enabling it to easily learn traversability in new circumstances. To this end, we first generate self-supervised traversability labels from past driving trajectories by labeling regions traversed by the vehicle as highly traversable. Using the self-supervised labels, we then train a neural network that identifies terrains that are safe to traverse from an image using a one-class classification algorithm. Additionally, we supplement the limitations of self-supervised labels by incorporating methods of self-supervised learning of visual representations. To conduct a comprehensive evaluation, we collect data in a variety of driving environments and perceptual conditions and show that our method produces reliable estimations in various environments. In addition, the experimental results validate that our method outperforms other self-supervised traversability estimation methods and achieves comparable performances with supervised learning methods trained on manually labeled data.