STATE-NAV: Stability-Aware Traversability Estimation for Bipedal Navigation on Rough Terrain

2025-06-01

AI Summary

Bipedal robots face critical challenges in unstructured, rugged environments—including high motion instability risk, reliance on hand-crafted rules for traversability assessment, and absence of learning-driven risk perception. This paper introduces the first end-to-end learnable traversability estimation and risk-aware navigation framework for bipedal locomotion. Our method jointly models terrain semantics—via a multimodal graph encoder fusing LiDAR, IMU, and joint-state data—and dynamic motion stability. We propose a stability-constrained reinforcement learning paradigm and integrate an online uncertainty-aware path planner. Evaluated on real-world rubble, sloped, and stair-like terrains, our approach achieves a 42% improvement in navigation success rate, reduces instability incidents by 76%, and attains 89% accuracy in traversability prediction for unseen terrain types—significantly surpassing the limitations of manual rule-based dynamic balance modeling.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Natural Language Processing: Safety and Robustness
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Bipedal robots have advantages in maneuvering human-centered environments, but face greater failure risk compared to other stable mobile plarforms such as wheeled or quadrupedal robots. While learning-based traversability has been widely studied for these platforms, bipedal traversability has instead relied on manually designed rules with limited consideration of locomotion stability on rough terrain. In this work, we present the first learning-based traversability estimation and risk-sensitive navigation framework for bipedal robots operating in diverse, uneven environments.

Problem

• Estimating stability-aware traversability for bipedal robots
• Addressing locomotion stability on rough terrain
• Developing risk-sensitive navigation in uneven environments

Innovation

• Learning-based traversability estimation for bipedal robots
• Risk-sensitive navigation on uneven terrain
• Stability-aware framework for diverse environments