Self-Supervised Traversability Learning with Online Prototype Adaptation for Off-Road Autonomous Driving

2025-04-16

AI Summary

Off-road autonomous driving demands real-time, robust terrain traversability analysis, yet faces critical challenges: absence of large-scale annotated datasets, severe onboard computational constraints, and stringent latency requirements (≥10 Hz). To address these, we propose the first self-supervised Bird’s-Eye-View (BEV) traversability assessment framework tailored for野外 environments. It pioneers the integration of BEV representation into self-supervised traversability learning; introduces a lightweight network architecture coupled with an online prototype clustering mechanism to enable dynamic, in-motion environment representation updates; and requires no manual annotations, producing traversability cost maps end-to-end. Evaluated on a diverse, real-world dataset spanning multiple seasons and geographic regions, our method significantly outperforms state-of-the-art approaches. Onboard deployment achieves 10 Hz inference throughput. A 5.5 km field experiment demonstrates that the generated cost maps directly enable downstream motion planning without further adaptation.

Technology Category

• Intelligent Robots: Localization, Mapping, and Navigation
• Computer Vision: Vision for Robotics & Autonomous Driving
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Achieving reliable and safe autonomous driving in off-road environments requires accurate and efficient terrain traversability analysis. However, this task faces several challenges, including the scarcity of large-scale datasets tailored for off-road scenarios, the high cost and potential errors of manual annotation, the stringent real-time requirements of motion planning, and the limited computational power of onboard units. To address these challenges, this paper proposes a novel traversability learning method that leverages self-supervised learning, eliminating the need for manual annotation. For the first time, a Birds-Eye View (BEV) representation is used as input, reducing computational burden and improving adaptability to downstream motion planning. During vehicle operation, the proposed method conducts online analysis of traversed regions and dynamically updates prototypes to adaptively assess the traversability of the current environment, effectively handling dynamic scene changes. We evaluate our approach against state-of-the-art benchmarks on both public datasets and our own dataset, covering diverse seasons and geographical locations. Experimental results demonstrate that our method significantly outperforms recent approaches. Additionally, real-world vehicle experiments show that our method operates at 10 Hz, meeting real-time requirements, while a 5.5 km autonomous driving experiment further validates the generated traversability cost maps compatibility with downstream motion planning.

Problem

• Self-supervised traversability learning for off-road autonomous driving
• Online prototype adaptation to handle dynamic scene changes
• BEV representation reduces computation and aids motion planning

Innovation

• Self-supervised learning eliminates manual annotation
• Birds-Eye View input reduces computational burden
• Online prototype adaptation handles dynamic scenes