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.
Updated
The defining link to reality
ISO 23247-1 defines a manufacturing digital twin around a fit-for-purpose representation and synchronisation with an observable physical element. Applied to robotics, the represented element might be one identified robot, its workcell or a deployed process. Sensor readings, calibration results, controller logs and maintenance records can update the representation at a cadence suited to its purpose.
The twin may combine a robot description, scene geometry, kinematics, dynamics, software state and learned models. It does not need to reproduce every property of the physical system. “Fit for purpose” means that a collision-checking twin and a predictive-maintenance twin can legitimately contain different information.
How it differs from a simulator
An ordinary simulator can model a generic robot or an imagined environment without corresponding to any particular physical instance. A digital twin has an identity relationship with a real counterpart and a maintained data path that keeps relevant properties aligned.
Simulation can therefore be one capability inside a twin, but simulation alone is insufficient. A static CAD model or one-off scene copy is better described as a digital model unless it is synchronised with the real asset for a defined use.
Robotics uses and limits
Digital twins can support calibration, monitoring, virtual commissioning, fault analysis and testing proposed controller changes. NIST's ISO 23247 use cases illustrate how operational data, modelling and simulation can be combined for machine health, scheduling and virtual commissioning.
Synchronisation does not guarantee physical fidelity. Unmodelled contact, wear, sensor noise or latency can still make predictions wrong, so claims about a twin should state what is updated, how often and against which real measurements it has been validated.
Sources
Related terms
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.
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.
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.
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.