Experience-Driven Continual Learning of Terrain Traversability for Quadruped Robots

2026-09-30

AI Summary

This study addresses the challenge that purely visual perception struggles to predict contact-related consequences, such as slippage and energy consumption, when quadruped robots navigate unknown terrains. To overcome this limitation, this work proposes a framework that extracts visual features using DINOv3 and employs an evidential regressor to predict proprioceptive metrics along with their uncertainties, thereby generating conservative traversability maps. Furthermore, a continual learning mechanism integrating bounded experience replay with validation gating is introduced to mitigate catastrophic forgetting during adaptation to novel terrains. The system is deployed on a Unitree Go2 platform within the ROS2 framework. Hardware experiments conducted across three unseen terrains demonstrate that, compared to a baseline without gated replay, the proposed method reduces final anchor negative log-likelihood degradation by 23.1%, effectively enhancing robust navigation capabilities in open-world environments.

Abstract

Safe and efficient quadruped navigation over unfamiliar terrain requires predicting terrain-robot interaction before contact: geometry and visual appearance alone cannot reveal how the robot will slip, load its feet, or expend energy. This paper presents a continual learning pipeline that uses locomotion experience to learn these interaction outcomes from pre-contact images and continually updates the predictions as new contacts are observed. Pre-contact descriptors, produced by a DINOv3 backbone model frozen during training, are mapped to five proprioceptive indicators weighted according to measurement reliability: planar foot slip, mean normal ground-reaction force, traction index, cost of transport, and touchdown loading rate. A compact evidential regressor allows us to predict these indicators together with aleatoric and epistemic uncertainty from the visual descriptors. Continual adaptation combines bounded experience replay with a validation gate: candidate models replace the deployed predictor only when they improve performance on recent held-out data while keeping degradation on historical held-out data within a prescribed tolerance. Predictions and epistemic uncertainty are projected into a local multilayer map and combined into a conservative traversability score map whose property weights can be adjusted without retraining. The resulting map is used for downstream navigation tests. The ROS2 implementation supports evaluation on a Unitree Go2 in simulation and on hardware, with models trained separately in each domain. On a sequential hardware stream over three previously unseen terrains, gated replay reduces final anchor negative log-likelihood (NLL) degradation by 23.1% relative to replay without the gate while attaining similar new-terrain adaptation.

Problem

• Quadruped Robots
• Terrain Traversability
• Continual Learning
• Terrain-Robot Interaction
• Uncertainty Estimation

Innovation

• Continual Learning
• Evidential Regression
• Terrain Traversability
• Experience Replay
• Uncertainty Estimation