ADAPT: Adaptive Dual-projection Architecture for Perceptive Traversal

TL;DR

ADAPT introduces an adaptive sensing architecture for humanoid robots that balances perception and efficiency by dynamically adjusting its environmental sensing range, leading to improved traversal performance in complex 3D environments.

Contribution

The paper presents ADAPT, a novel dual-projection sensing architecture with learnable sensing range, enabling efficient and robust humanoid traversal in diverse environments.

Findings

Reduces observation dimensionality and computational overhead.

Accelerates training compared to voxel-based baselines.

Achieves successful zero-shot transfer and robust traversal performance.

Abstract

Agile humanoid locomotion in complex 3D en- vironments requires balancing perceptual fidelity with com- putational efficiency, yet existing methods typically rely on rigid sensing configurations. We propose ADAPT (Adaptive dual-projection architecture for perceptive traversal), which represents the environment using a horizontal elevation map for terrain geometry and a vertical distance map for traversable- space constraints. ADAPT further treats its spatial sensing range as a learnable action, enabling the policy to expand its perceptual horizon during fast motion and contract it in cluttered scenes for finer local resolution. Compared with voxel-based baselines, ADAPT drastically reduces observation dimensionality and computational overhead while substantially accelerating training. Experimentally, it achieves successful zero-shot transfer to a Unitree G1 Humanoid and signifi- cantly outperforms fixed-range baselines, yielding highly robust traversal across diverse 3D environtmental challenges.