What is whole-body control in robotics?
Whole-body control coordinates every joint — legs, torso, arms, head — as one system so the robot stays balanced while reaching, ducking or carrying.
Updated:
Definition
Whole-body control treats the robot as one kinematic and dynamic chain. Controllers allocate torques or joint targets across the full body so balance, contact and task goals are satisfied together.
How it works
A high-level planner (or learned policy) proposes a short motion; a tracker runs at tens of hertz with feedback from IMUs, joint encoders and often lidar or cameras. Priorities (don't fall, then clear obstacles, then reach the goal) are encoded as weighted tasks or constraints.
Why it matters
Homes and factories are built for people. A robot that can only walk in open space fails when it must duck under a beam or turn sideways through a door. Whole-body control is the difference between a demo on a clear floor and useful motion in clutter.
Frequently asked questions
How is it different from walking-only control?
Walking stacks often ignore arms and torso clearance. Whole-body control plans feet and upper body together so shoulders clear doorways while feet stay stable.
Where is it used?
Humanoids navigating clutter, mobile manipulators that roll and reach at once, and research frameworks like Galbot PASSAGE.
Does it need torque sensors?
Often yes at key joints, plus state estimation — but some learned controllers approximate contact forces from kinematics and vision.
Coverage
Galbot's PASSAGE teaches humanoids to duck, step over and squeeze through clutter
On a Unitree G1, Galbot's planner–tracker stack finished 50 unseen physical layouts without falls; 45 were contact-free. Destination success in sim was 98.67% — contact-free 70.3%.
Newsletter
Robrief Weekly
One email a week with the robotics news that matters. Free, unsubscribe anytime.