Robot control

Bilateral teleoperation

Definition

Bilateral teleoperation exchanges commands from an operator-side device to a remote robot and feedback from the robot or environment back to the operator. The returning signal often represents contact force so the operator can feel part of the remote interaction.

Also known as: Force-reflecting teleoperation

Updated

Close a loop through the operator and environment

In ordinary one-way teleoperation, a leader device may send positions to a follower robot while video provides separate feedback. A bilateral system couples motion and interaction in both directions. When the follower touches an object, measured or estimated force can be reflected at the leader device.

This coupling can help an operator judge contact in insertion, surface following, or robotic manipulation. It also means the human, both robots, the environment, and the communication channel all participate in one feedback loop.

Transparency competes with stability

Transparency describes how faithfully the remote environment feels at the operator side. Strong, immediate force reflection can improve that fidelity, but delay, sampling, scaling, and uncertain dynamics can destabilize the coupled system. The classic Lawrence analysis formalizes the relationship between stability and transparency in bilateral teleoperation.

Controllers may use passivity, damping, wave variables, or limited feedback bandwidth to preserve stability. These measures can make contact feel softer or more delayed than it is. Stable operation therefore does not imply perfect force reproduction.

Force can be measured or estimated

A force-torque sensor can provide the follower-side interaction signal, but cost and packaging may make direct sensing difficult. Yamane and colleagues report a four-channel system that estimates velocity and external force with disturbance observers on low-cost manipulators. Their real-robot and imitation-learning results apply to that hardware, estimator, and tested tasks, not to every sensorless system.

Network delay, packet loss, actuator saturation, calibration error, and unexpected hard contact remain important limits. Evaluation should distinguish force reflection from visual feedback, state the communication conditions, and report both contact fidelity and closed-loop stability.

Sources