Collision-Free Humanoid Traversal in Cluttered Indoor Scenes

TL;DR

The paper tackles collision-free, whole-body humanoid traversal in cluttered indoor environments by introducing HumanoidPF, a differentiable potential-field-inspired perceptual representation that encodes humanoid–obstacle relationships to provide dense guidance for reinforcement learning. It pairs HumanoidPF with a hybrid scene generation pipeline and a specialist-to-generalist training regime to achieve robust generalization across diverse cluttered scenes, while enabling real-world deployment through a Click-and-Traverse teleoperation system. Central contributions include the geodesic-distance-based attractive and distance-based repulsive fields, a priority-weighting scheme for multi-joint coherence, and a von Mises–Fisher motion-prior reward that aligns policy actions with the field; the approach demonstrates strong sim-to-real transfer and practical applicability in real indoor environments. Overall, the method achieves higher traversal success and robustness than baselines, generalizes to unseen clutter, and enables intuitive teleoperation for domestic humanoids in cluttered scenes.

Abstract

We study the problem of collision-free humanoid traversal in cluttered indoor scenes, such as hurdling over objects scattered on the floor, crouching under low-hanging obstacles, or squeezing through narrow passages. To achieve this goal, the humanoid needs to map its perception of surrounding obstacles with diverse spatial layouts and geometries to the corresponding traversal skills. However, the lack of an effective representation that captures humanoid-obstacle relationships during collision avoidance makes directly learning such mappings difficult. We therefore propose Humanoid Potential Field (HumanoidPF), which encodes these relationships as collision-free motion directions, significantly facilitating RL-based traversal skill learning. We also find that HumanoidPF exhibits a surprisingly negligible sim-to-real gap as a perceptual representation. To further enable generalizable traversal skills through diverse and challenging cluttered indoor scenes, we further propose a hybrid scene generation method, incorporating crops of realistic 3D indoor scenes and procedurally synthesized obstacles. We successfully transfer our policy to the real world and develop a teleoperation system where users could command the humanoid to traverse in cluttered indoor scenes with just a single click. Extensive experiments are conducted in both simulation and the real world to validate the effectiveness of our method. Demos and code can be found in our website: https://axian12138.github.io/CAT/.

Figures (4)

• Figure 1: Using a single generalist policy, our humanoid robot achieves collision-free traversal in cluttered indoor environments, including (a) detouring through narrow passages, (b) crouching under low-hanging obstacles, (c) squeezing through tight indoor spaces, and (d) hurdling over objects scattered on the floor.
• Figure 2: Overall pipeline. We learn a visuomotor policy that maps diverse obstacle geometries and spatial layouts to corresponding whole-body traversal skills. Left: HumanoidPF for whole-body traversal learning.(Top) Construction of HumanoidPF, a reformulation of APF tailored for humanoid whole-body traversal; (Bottom) its use as informative perceptual representation and collision-avoidance rewards. Right: Scalable training and deployment pipeline.(Top) Hybrid scene generation for constructing diverse and challenging training environments; (Middle) parallel training of multiple specialist policies followed by distillation into a single generalist policy; (Bottom) sim-to-real deployment via "Click-and-Traverse", an intuitive loco-navigation teleoperation in cluttered indoor scenes. Sections \ref{['sec:apf_humanoid']}, \ref{['sec:generalize']} and \ref{['sec:click_and_traverse']} provide detailed descriptions of the HumanoidPF for traversal learning, the scalable training, and deployment pipeline, respectively.
• Figure 3: (a) Construction of the APF and (b) motion prior distribution induced by the HumanoidPF.
• Figure 4: Collision-free humanoid traversal in both simulation and the real world.(a) Humanoid traversal behaviors on eight representative test scene types; (b) traversal behaviors in procedurally generated cluttered environments; (c) real-world "hurdle-crouch" scenario, validating sim-to-real transfer in a cluttered indoor setting; (d) robustness under dynamic disturbances, where simple object movements (blue arrows) are introduced during traversal.