Data & collection
Teleoperation
Teleoperation is real-time human control of a robot from a remote or mediated interface. For humanoid training-data collection, the operator’s inputs are mapped to robot motions while cameras, proprioception, commands and outcomes are recorded, producing embodied demonstrations in the robot’s own observation and action spaces.
Also known as: tele-operation, remote operation
Updated
From operator motion to robot action
ISO 8373:2021 defines teleoperation as real-time human control of robot motion from a remote site. The interface can be a joystick, a pair of leader arms, a VR system or a pose-tracking camera. “Real-time” means that the operator acts through a live feedback loop; it does not imply zero delay or a direct one-to-one mapping.
The mapping depends on the embodiment. In ALOHA, backdriven leader arms command bimanual follower arms. A humanoid system may instead convert a person’s hand or body pose into feasible targets, while a lower-level controller maintains balance and respects joint limits. OmniH2O uses kinematic pose as an interface for whole-body humanoid control through VR or RGB-camera input.
Why teleoperation produces useful training data
Teleoperation records the task from the robot’s own sensors while preserving the actions that caused each change. That makes the result more directly useful for policy learning than ordinary human video, which usually lacks robot commands and proprioception. A collection can include multiple camera streams, joint state, operator input, executed action, force signals, language and success labels.
The method does not guarantee clean demonstrations. Network delay, poor ergonomics, action scaling and operator corrections all appear in the data. The DROID collection used a shared teleoperation and robot stack to improve consistency across sites. For humanoids, collectors should also record controller mode, safety interventions and whether commanded motion was clipped, because the operator’s request and the robot’s executed action may differ.
Sources
Related terms
Data & collection
Demonstration
A demonstration is a recorded example of how an intended task or behaviour is performed, usually represented as a time-aligned sequence of observations, states and actions. For humanoid robot learning, demonstrations may come from teleoperation, kinaesthetic guidance, motion capture or autonomous experts and provide targets for imitation.
Data & collection
Trajectory
A trajectory is a time-ordered sequence of states or observations, actions and, where applicable, rewards generated as an agent or robot evolves. A complete episode or policy rollout often yields a trajectory, but the terms are not universally identical: trajectories may be partial, while episodes have dataset- or environment-defined boundaries.
Hardware & control
Whole-body control
Whole-body control coordinates a humanoid’s full multibody state, actuated joints and environmental contacts to pursue several motion or force objectives while respecting constraints such as balance, joint limits and friction. It is a family of hierarchical or optimisation-based methods, not one algorithm; implementations may output joint positions, accelerations or torques.
Hardware & control
Proprioception
Proprioception is sensing of a robot’s own internal configuration and motion rather than the external scene. For a humanoid it commonly includes joint positions and velocities, actuator effort or torque, and inertial measurements of body rotation and acceleration. These signals support state estimation and feedback control but do not, by themselves, directly describe nearby objects or terrain.