Ego-Vision World Model for Humanoid Contact Planning
By: Hang Liu , Yuman Gao , Sangli Teng and more
Potential Business Impact:
Robots learn to touch and use things safely.
Enabling humanoid robots to exploit physical contact, rather than simply avoid collisions, is crucial for autonomy in unstructured environments. Traditional optimization-based planners struggle with contact complexity, while on-policy reinforcement learning (RL) is sample-inefficient and has limited multi-task ability. We propose a framework combining a learned world model with sampling-based Model Predictive Control (MPC), trained on a demonstration-free offline dataset to predict future outcomes in a compressed latent space. To address sparse contact rewards and sensor noise, the MPC uses a learned surrogate value function for dense, robust planning. Our single, scalable model supports contact-aware tasks, including wall support after perturbation, blocking incoming objects, and traversing height-limited arches, with improved data efficiency and multi-task capability over on-policy RL. Deployed on a physical humanoid, our system achieves robust, real-time contact planning from proprioception and ego-centric depth images. Website: https://ego-vcp.github.io/
Similar Papers
ContactRL: Safe Reinforcement Learning based Motion Planning for Contact based Human Robot Collaboration
Robotics
Robots learn to touch people safely during work.
Learning to Act Through Contact: A Unified View of Multi-Task Robot Learning
Robotics
Robot learns many jobs with one brain.
Contact-Safe Reinforcement Learning with ProMP Reparameterization and Energy Awareness
Robotics
Robots learn to move safely and smoothly.