Humanoids move naturally and flexibly using attention to plan contacts
Humanoid Mobility via Attention-based Contacts
Robotics
Summary
Walking or moving like humans across tricky ground is hard because robots need to place their feet just right and move their whole body smoothly. The authors created a system that learns how to predict where a robot should touch the ground next, then moves the robot's body to follow that plan. This method uses examples of real motions to guide learning but doesn’t copy them exactly, giving the robot flexibility to handle new situations. The system works using only what the robot senses itself and can adapt to different terrains without needing exact motion instructions.
What this means in practice
- •For robotics engineers: Develop robots that walk naturally across uneven or complex surfaces by using learned contact plans combined with adaptable whole-body control.
- •For video game developers: Create realistic humanoid character movements in games that adapt to dynamic terrain by using the contact-aware motion planning approach.
Authors
Junzhe He, Kehan Wen, Chong Zhang, Clemens Schwarke, Zewei Zhang, Kaixian Qu, Siyuan Hu, Matthias Müller, Vladlen Koltun, Marco Hutter
Abstract
Humanoid mobility over complex terrain requires accurate foot placement and coordinated whole-body motion capable of robust adaptation. Reinforcement learning (RL) policies trained without motion references generalize well across terrains, yet eliciting effective whole-body coordination through reward design and exploration alone remains challenging. In contrast, reference-based methods yield natural motions but struggle to adapt beyond the predefined trajectories in their training data. We introduce a learning framework that bridges these two approaches by using motion data strictly as a training prior. An explicit, self-supervised, attention-based contact planner predicts future body-environment contacts, while a contact-conditioned tracker translates these plans into coordinated whole-body motion. Together, they establish contact as an interface between sensory observations and control actions. This abstraction allows the policy to internalize the contact-aware structure of locomotion rather than directly replicating kinematics. Consequently, a single policy can exploit data-driven motion priors for natural whole-body coordination while retaining the flexibility to discover environment-dependent behaviors through reinforcement learning. Building on the learned contact planner and tracker, a subsequent fine-tuning stage bootstraps their acquired capabilities to explore environment variations beyond those covered by the original motion data. Operating entirely from onboard observations without any reference inputs at deployment, this framework enables versatile humanoid mobility in challenging environments.