Humanoid robot learns to balance and control force with hands
Learning Force-Reactive Humanoid Loco-Manipulation
RoboticsArtificial IntelligenceMachine Learning
Summary
Many everyday tasks need not just movement but also the right amount of force when touching or holding things. The authors designed a system called FLORE that teaches humanoid robots to use their hands and body to control forces while moving around. They demonstrated their approach by collecting human-like demonstrations without using robots directly, then trained a model to predict both movements and forces. In tests, their robot was much better at tasks requiring careful force control compared to methods that only controlled movement.
What this means in practice
- •For robotics engineers: Build humanoid robots that handle objects delicately while walking or balancing using force-reactive control.
- •For industrial automation teams: Improve robots that manipulate tools and materials in contact-rich environments with both motion and force coordination.
Authors
Sihang Li, Xiaofeng Guo, Jinkun Cao, Charlie Cheng, Ruihan Yang, Chen Feng, Guanya Shi, Haozhi Qi
Abstract
Many everyday tasks require controlling contact force as well as motion. On a humanoid, this force must often be coordinated across both hands and the whole body, and cannot be specified by kinematic trajectories alone. Therefore, we present FLORE, a framework for learning force-reactive humanoid loco-manipulation. We collect robot-free demonstrations using two force-sensing UMI-style handheld grippers for hand motion and contact wrenches, and wearable trackers for body motion. A force-reactive diffusion policy takes vision, proprioception, and measured wrenches as input, then predicts actions and per-hand force commands. A hybrid compliance whole-body controller with force tracking executes these commands while maintaining balance: an integral of the force error measured by real force/torque (F/T) sensors lets each hand track force accurately along the commanded direction and remain compliant along the others. In simulation, the controller tracks per-hand force commands of varied magnitude and direction, with little loss in whole-body motion accuracy. Across three contact-rich tasks on the robot, the force-reactive policy succeeds in 80% of rollouts, compared with 26.7% for a position-only policy trained on the same demonstrations and executed by the same controller. For more details, please visit our website: https://louis-leee.github.io/FLoRe/.