Reference
Humanoid Robotics Glossary
Clear, source-backed definitions of humanoid robotics, embodied AI, robot learning, training data, control, and simulation terms.
A
Models & learning
Action chunking
Action chunking is a policy design in which one observation is used to predict a sequence of several future robot actions rather than a single next action. The chunk can capture short-term motion structure and reduce the effective number of sequential predictions, while its execution and replanning schedule determine how quickly the robot can react to new observations.
Hardware & control
Actuator
An actuator is a robot’s power mechanism that converts electrical, hydraulic, pneumatic or other supplied energy into controlled mechanical force, torque or motion. In a humanoid, actuators drive joints, often through transmissions; they are distinct from the joints and sensors, and their arrangement need not correspond one-to-one with the robot’s degrees of freedom.
Models & learning
Asynchronous inference
Asynchronous inference is a deployment pattern that overlaps robot action execution with prediction of a later action chunk. A robot client consumes queued actions while a policy process or server evaluates a newer observation, then merges or replaces the remaining queue when the new chunk arrives. The pattern reduces idle inference gaps but does not remove latency, stale observations or safety constraints.
B
C
D
Hardware & control
Degree of freedom
A degree of freedom (DoF) is one independent parameter needed to specify a robot’s configuration; equivalently, a robot’s DoF is the dimension of its configuration space. It describes possible motion, not the number of motors. Joint constraints, closed kinematic chains and environmental contacts can make joint count, actuator count and controllable motion differ.
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.
Models & learning
Diffusion policy
A diffusion policy is a learned robot policy that generates an action or action sequence through iterative denoising conditioned on observations and, where applicable, a task goal. This formulation can represent several plausible behaviours for the same scene and is commonly trained by imitation on time-aligned trajectories of observations and expert actions.
Simulation & transfer
Digital twin
A digital twin is a fit-for-purpose digital representation of a specific physical robot, asset or process that is kept synchronised with its real counterpart through operational data. It may contain geometry, dynamics and simulation models, but the maintained link to an identified real system distinguishes it from an ordinary, standalone simulator.
Simulation & transfer
Domain randomisation
Domain randomisation is a simulation-training technique that varies selected visual, physical or sensor parameters across generated examples or episodes. By exposing a model or policy to a distribution of environments rather than one calibrated scene, it aims to make real conditions fall within the learned variation and improve sim-to-real transfer.
E
Data & collection
Egocentric data
Egocentric data is sensor data recorded from the viewpoint of the person or robot performing an activity, most commonly with a head- or body-mounted camera. It can also include audio, gaze, depth or inertial signals. For humanoid learning, it shows hands, objects and actions from an actor-centred perspective.
Hardware & control
End effector
An end effector is a task-specific device attached to a robot manipulator’s mechanical interface so the robot can act on its environment, such as a gripper, hand or welding tool. It is distinct from the wrist or mounting flange, and from the tool centre point, which is only a coordinate frame used to plan the device’s motion.
F
Hardware & control
Feedback control
Feedback control is a closed-loop control method that measures a system’s current output or state, compares it with a target and adjusts the command using the resulting error. In robotics, feedback can correct joint, end-effector, balance or force errors as new sensor measurements arrive.
Models & learning
Flow matching
Flow matching is a generative-model training method that learns a time-dependent vector field transporting samples from a simple prior distribution to a data distribution. Training regresses the vector field along chosen probability paths without solving the flow during each update; generation starts from the prior and numerically integrates the learned ordinary differential equation.
I
Models & learning
Imitation learning
Imitation learning is a family of methods that learns a policy from examples of expert behaviour rather than specifying every control rule by hand. In robotics, demonstrations pair observations or states with actions, trajectories or inferred objectives. Behaviour cloning is one imitation-learning method; interactive and inverse approaches address different supervision and distribution-shift problems.
Hardware & control
Impedance control
Impedance control regulates the dynamic relationship between a robot’s motion error and the force or torque it exerts, usually by giving the robot virtual stiffness, damping and sometimes inertia. It lets an end effector or limb yield predictably during contact instead of trying to hold an exact position regardless of interaction forces.
M
Models & learning
Markov decision process
A Markov decision process is a mathematical model of sequential decision-making defined by states, actions, transition probabilities and rewards. After an agent chooses an action, the current state and action determine the distribution of the next state and reward. Reinforcement-learning methods use this structure to compare policies by expected cumulative reward.
Data & collection
Motion capture
Motion capture is the measurement and reconstruction of a person’s or object’s movement over time, commonly as joint positions, orientations or a fitted body model. Optical markers, cameras and inertial sensors can supply the measurements. Humanoid robotics uses the resulting motion sequences for analysis, imitation and retargeting to a robot body.
Simulation & transfer
Motion retargeting
Motion retargeting is the adaptation of a recorded or generated motion from one body to another with different proportions, joints or limits. For humanoid robots, it maps source poses or trajectories into robot configurations while preserving task-relevant relationships such as contacts and end-effector paths and satisfying kinematic, balance, collision and actuator constraints.
P
Models & learning
Policy
A policy is the decision rule that maps a robot’s current observations or estimated state, and sometimes a task instruction, to an action or probability distribution over actions. It can be hand-designed or learned from demonstrations, rewards or both. In humanoid robotics, its outputs may be joint targets, torques, end-effector changes or higher-level skills.
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.
R
Simulation & transfer
Real-to-sim
Real-to-sim is the process of using observations or measurements from a real robot, object or environment to construct or update a simulation-ready representation. Depending on the application, this may recover geometry, articulation, calibration and physical parameters, or translate real demonstrations into simulator state, usually to support later analysis, training or sim-to-real deployment.
Models & learning
Reinforcement learning
Reinforcement learning is a method in which an agent learns a policy by interacting with an environment and optimising cumulative reward. In humanoid robotics, actions change the robot and world, while observations, rewards and episode endings provide experience for improving balance, locomotion or manipulation behaviour.
Models & learning
Robot foundation model
A robot foundation model is a broadly pretrained model intended to provide a reusable starting point for multiple robot tasks, environments or embodiments. It learns from diverse robotics and sometimes web or human data, then acts directly or is adapted with target-domain data. The term describes a training and reuse strategy, not one fixed architecture.
Hardware & control
Robot kinematics
Robot kinematics describes the geometric relationship between a robot’s joint configuration and the position, orientation and velocity of its links or end-effector, without modelling the forces that cause the motion. Forward kinematics computes pose from joint values; inverse kinematics searches for joint values that achieve a requested pose.
Models & learning
Robot learning
Robot learning is the use of data and machine-learning methods to acquire or improve robot perception, prediction, planning or control. A robot may learn from demonstrations, rewards, direct interaction, simulation or previously collected datasets. The field includes reinforcement learning and imitation learning, but neither term is a synonym for robot learning as a whole.
S
Simulation & transfer
Sim-to-real
Sim-to-real is the transfer of a model, policy or behaviour developed wholly or partly in simulation to a physical robot or real environment. The central problem is the reality gap: errors in simulated dynamics, sensing, appearance and timing can make a strategy successful in simulation but unreliable or unsafe on hardware.
Simulation & transfer
Synthetic data
Synthetic data is training or evaluation data produced computationally rather than measured directly from the target real-world event. In robotics it can include rendered sensor observations, automatically generated labels, reconstructed or generated scenes, and simulated trajectories. A simulated rollout is one subtype: a temporally linked trajectory produced by executing a controller or policy in a simulator.
T
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.
Hardware & control
Torque control
Torque control is a control mode in which desired joint torques, or equivalent linear forces, are the command variables sent to a robot’s low-level actuators. Unlike position control, it does not directly command where each joint should be. Motion and contact instead emerge from the applied efforts, robot dynamics, feedback, constraints and any higher-level controller.
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.
V
W
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.
Models & learning
World model
A world model is a learned predictive model of how an environment changes, often conditioned on a robot’s current state and candidate actions. It may forecast pixels, latent states, rewards or other task-relevant quantities. A policy or planner can use those predictions to compare possible futures before commanding the physical robot.
Models & learning
World-action model
A world-action model is an emerging, non-standardised name for a robot model that couples prediction of future world states with generation of executable actions, often by adapting a video or world-model backbone as a policy. It differs from an action-conditioned world model that only predicts consequences of supplied actions, and from Barrett’s WAM, the Whole Arm Manipulator.