Humanoid robots learn to move safely through cluttered spaces
PASSAGE: Scaling Scene-Aligned Motion Learning for Perceptive Humanoid Traversal in Cluttered Environments
Robotics
Summary
Humanoid robots can step over and squeeze past obstacles, but getting them to decide how to move just from what they see is hard. The authors created a system named PASSAGE that learns from lots of human movement data in cluttered rooms using virtual reality. It plans short moves based on what the robot sees and remembers, then controls the robot's whole body to safely navigate around obstacles in real time. Their approach improves with more data and works on real robots without needing maps or external computers.
What this means in practice
- •For robotics engineers: Implement onboard planning and control for humanoid robots to navigate cluttered environments without prebuilt maps using egocentric perception.
- •For warehouse automation teams: Deploy humanoid robots capable of adaptive traversal in cluttered storage areas, improving flexibility and safety over fixed-path navigation.$Commercial implications: Enables sale of adaptive humanoid robots for dynamic warehouse environments needing flexible obstacle traversal.
Authors
Yuxuan Ma, Zicheng Zeng, Chunlin Peng, Zhoujian Li, Zetong Zhao, Zhikai Zhang, Yunrui Lian, Han Xue, Sikai Liang, Weiyi Zhu, Mulin Chen, Chenghuai Lin, Jiayu Zeng, Yanwei An, Songan Zhang, Jiayuan Gu, Jilong Wang, Jingbo Wang, He Wang, Li Yi
Abstract
Humanoid robots can step over, squeeze past, and duck under obstacles, but learning to select and coordinate these behaviors from onboard perception remains challenging. Many existing approaches rely on task-specific reinforcement-learning objectives or curated motion libraries, making broad behavioral coverage costly. We present PASSAGE, a perception-conditioned planner--tracker framework for humanoid traversal. Using virtual reality and inertial motion capture, we collect 100 h of scene-aligned human motion across 1,500 cluttered scenes. A conditional flow-matching planner generates short-horizon references from motion history, a local destination, and a robot-centric multi-layer elevation map, while a perceptive whole-body tracker executes them at 50 Hz with geometric feedback. Real-time chunking promotes inter-chunk consistency, and planner-side RL post-training under the frozen tracker further improves closed-loop performance. Without skill annotations or obstacle-specific policies, one planner--tracker pair selects and composes traversal behaviors across unseen geometries. In simulation, component ablations quantify the contribution of each stage. Across three independent training seeds, scaling captured data from 6 to 100 h increases mean contact-free success from 48.1% to 68.9% on held-out scenes, while the final model with validated scene augmentation reaches 70.3%. The fully onboard system integrates egocentric 3D LiDAR perception, online occupancy mapping, 6.25 Hz planning, and 50 Hz control on a Jetson AGX Orin; tests across 50 unseen physical layouts demonstrate traversal without prebuilt maps or offboard computation.