Admittance Control

Measures the external force applied, computes the resulting motion from a set mass-damper-spring relationship, and hands that to the position loop.

Admittance control is an interaction-control method that makes a robot yield to external force, and it's theoretically grounded in the duality between impedance and admittance from Hogan's 1985 paper on impedance control. Its input is force and its output is motion: a wrist-mounted six-axis force-torque sensor measures the external force F, and the controller computes how the robot should move from F = M·a + D·v + K·(x − x_d) (M, D, K are the virtual mass, damping, and stiffness, and x_d is the originally intended target position), then hands the corrected position or velocity command off to the robot's own built-in position loop. Impedance control runs in the opposite direction: it takes a position deviation as input and outputs a force or torque, which requires the joints to be able to control torque directly. Admittance control is therefore well suited to industrial arms that only expose a position interface, have high gear ratios, and are inherently stiff — but it can oscillate against very stiff environments. ROS 2's ros2_controllers package includes an admittance_controller, which can be used for zero-force dragging.

ExampleMount a six-axis force-torque sensor on an industrial arm's end effector, turn on admittance control, and set the stiffness K to zero: pushing the end effector by hand moves the robot along with the push, and it settles to a stop under the damping term once released — one way of implementing hand-guided teaching.

Also called

Admittance Controller

Related

Impedance Control · Compliance Control · Six-Axis Force/Torque Sensor · Zero-Force Drag · Force Control · Position Control

Sources

ros2_controllers: Admittance Controller
Wikipedia: Impedance control
Hogan, Impedance Control: An Approach to Manipulation, Part I (1985)