Task-Conditioned Uncertainty Costmaps for Legged Locomotion

2026-04-30

AI Summary

This work addresses the challenge of accurate foothold prediction in highly unstructured terrain, where existing methods often fail to ensure motion planning feasibility due to unreliable perception inputs. The authors propose a task-conditioned cognitive uncertainty modeling approach that integrates foothold prediction, out-of-distribution (OOD) detection, and uncertainty estimation within a unified framework to generate an uncertainty-aware cost map. This method effectively identifies OOD regions and quantifies uncertainty arising from perceptual deficiencies, even under limited training data. Experimental results demonstrate significant improvements in OOD detection performance in both simulation and real-world environments, with up to a 37% reduction in planning feasibility error compared to geometry-only baseline methods, thereby exhibiting superior reliability and robustness.

Abstract

Legged robots maintain dynamic feasibility through multicontact interactions with terrain. Learned foothold prediction can provide feasibility-aware costs for motion planning and path selection, but accurately predicting future contacts from perceptual inputs such as height scans remains challenging on highly unstructured terrain, even with a repetitive gait cycle. In this work, we show that modeling epistemic uncertainty in predicted footholds, conditioned on terrain observations and commanded motion, distinguishes in-distribution from out-of-distribution operating regimes in simulation and real-world settings. This allows a single learned model, trained on limited data distributions, to express uncertainty caused by missing training coverage. We use this learned uncertainty to detect OOD regions and incorporate them into a unified costmap-generation framework for uncertainty-aware path planning. Using these uncertainty-aware costmaps, we evaluate feasibility error across in-distribution and OOD terrains in simulation and real-world settings. The results show improved OOD detection, up to a 37% reduction in simulation feasibility error, and more reliable planning behavior than geometry-only baselines.

Problem

• legged locomotion
• foothold prediction
• epistemic uncertainty
• out-of-distribution detection
• terrain perception

Innovation

• epistemic uncertainty
• task-conditioned
• uncertainty-aware costmaps
• legged locomotion
• out-of-distribution detection