REDACT: Robust Perceptive Locomotion under Unseen Visual Corruption

Natapat Kirdwichai, Tobias Driskell-Poole, Andrei Sontea, Jadu Dash, Muhammad Burhan Hafez, Danesh Tarapore · N/A · 2026

Summary

Depth-conditioned locomotion policies have demonstrated impressive agile maneuvers, but can be steered to unpredictable actions when observations are outside their training distribution.

Abstract Summary

Depth-conditioned locomotion policies have demonstrated impressive agile maneuvers, but can be steered to unpredictable actions when observations are outside their training distribution. Occlusion, invalid returns, sensor noise, and visual distractors can shift deployment observations away from nominal simulated depth. While synthetic sensor augmentation targets specified degradations, it does not by itself define behavior under corruption families omitted from training. To address gaps in training-time coverage, we present REDACT (Retaining Evidence Despite Artifacts for Continued Traversal), a teacher-student framework combining an improved visual encoder architecture, persistent feature masking, and a novel consensus-gating algorithm to retain useful depth information under unmodeled corruption. The gate uses approximate conformal calibration on clean observations alone, requiring no prior knowledge of the corruption type. Trained on clean simulated depth, REDACT retains useful visual information under unseen corruption, supporting higher traversal success than existing parkour baselines. Evaluation of depth augmentation across corruption families further shows that REDACT improves robustness where augmentation coverage is missing. Real-world trials demonstrate zero-shot transfer to structured and forested environments with unfamiliar scene content.

Key Points

Depth-conditioned locomotion policies have demonstrated impressive agile maneuvers, but can be steered to unpredictable actions when observations are outside their training distribution.
Occlusion, invalid returns, sensor noise, and visual distractors can shift deployment observations away from nominal simulated depth.
While synthetic sensor augmentation targets specified degradations, it does not by itself define behavior under corruption families omitted from training.
To address gaps in training-time coverage, we present REDACT (Retaining Evidence Despite Artifacts for Continued Traversal), a teacher-student framework combining an improved visual encoder architecture, persistent feature masking, and a novel consensus-gating algorithm to retain useful depth information under unmodeled corruption.
The gate uses approximate conformal calibration on clean observations alone, requiring no prior knowledge of the corruption type.

Method

We train a privileged teacher and a depth-conditioned student to retain terrain information needed for locomotion while limiting the influence of unseen visual corruption (Fig. 3 ). Regularized multi-query cross-attention first structures the teacher’s terrain representation to reduce dependence on privileged detail that onboard depth cannot recover.

The teacher then initializes a depth-conditioned student whose residual encoder is designed to limit sensitivity to visual perturbations while preserving spatial evidence for consensus gating.

Method — REDACT: Robust Perceptive Locomotion under Unseen Visual Corruption — Fig. 4: Simulated training terrains: tilted ramps, gaps, hurdles, and steps.

Results

We evaluate whether REDACT maintains successful traversal when depth observations differ from those encountered during training. Comparisons with parkour baselines and augmentation-trained policies assess robustness beyond training coverage. We then examine the contributions of the encoder and selective rejection through component ablations and gate analysis.

Real-world trials assess zero-shot transfer to structured and forest environments. Training and Preprocessing.

Results — REDACT: Robust Perceptive Locomotion under Unseen Visual Corruption — Fig. 8: Depth decoding across three frames of simulated vine occlusion. Columns show clean depth, corrupted input, clean-input decoding, ungated and gated corrupted-input decoding, and flagged cells. $L_{1}$ errors reference the clean-input reconstruction. Flag counts appear below the images.

Sources and Demonstrations

Figures are reproduced from Kirdwichai et al. See the full paper for experimental details and the project page, when the paper links one, for demonstrations and videos.