Simulation
Choose a simulator for the project's hardest requirement, then define the
evidence that would make its results trustworthy.
Pick by bottleneck
- For a new robot app or first simulation demo, use
architect to check compatible reference apps before
selecting a fresh stack. Keep a selected example's simulator. Continue here
when no example fits or simulator choice is the actual unresolved question;
comparisons and existing simulation work do not need onboarding.
- Gazebo: prefer for ROS-centric mobile robots, standard sensor simulation,
and projects without a confirmed compatible NVIDIA GPU.
- Isaac Sim: consider when photorealism, synthetic data at scale, or the
NVIDIA Isaac stack is central. Confirm the current RTX GPU, driver, OS, and
VRAM requirements before committing to it; use
gazebo if that environment
is not real and available.
- MuJoCo: prefer for lightweight, contact-rich manipulation that does not
need ROS integration or photoreal rendering.
- A project may use different simulators for different concerns. Avoid forcing
one tool to own both navigation validation and training-data generation.
After selection, use gazebo, isaac-sim, or mujoco for tool mechanics.
Use isaac-lab for training workflows on top of Isaac Sim.
Define fidelity before building
- Name what must transfer to hardware: interfaces, sensor behavior, dynamics,
contacts, timing, or visual appearance.
- Match real sensor rate, noise, frame names, range, field of view, resolution,
and timestamp source where those affect downstream decisions.
- Record known model gaps instead of treating simulation success as real-world
evidence.
- If the simulator will back regression tests, control time step and randomness
and judge behavior with tolerances rather than pixel-perfect coincidence.
Read sensor fidelity when adding sensors, preparing
sim-to-real work, or turning a scenario into a repeatable test. Use data when
the unresolved question is what training data to generate rather than how to
simulate it.
Done
- The selected tool fits the available hardware and the project's main
fidelity requirement.
- The simulation contract names which interfaces and physical properties must
match reality.
- Every consumer uses a consistent clock and frame convention.
- Repeatability is designed where the simulation will be used as a test.
- Version and hardware requirements were checked in the current upstream
Gazebo,
Isaac Sim, or
MuJoCo documentation.
1---2name: simulation3description: Choose between Gazebo, Isaac Sim, and MuJoCo and define what the simulation must prove.4metadata:5 source: https://skillmd.com/skills/robium-ai/simulation6---78# Simulation910Choose a simulator for the project's hardest requirement, then define the11evidence that would make its results trustworthy.1213## Pick by bottleneck1415- For a new robot app or first simulation demo, use16 [architect](../architect/SKILL.md) to check compatible reference apps before17 selecting a fresh stack. Keep a selected example's simulator. Continue here18 when no example fits or simulator choice is the actual unresolved question;19 comparisons and existing simulation work do not need onboarding.20- **Gazebo:** prefer for ROS-centric mobile robots, standard sensor simulation,21 and projects without a confirmed compatible NVIDIA GPU.22- **Isaac Sim:** consider when photorealism, synthetic data at scale, or the23 NVIDIA Isaac stack is central. Confirm the current RTX GPU, driver, OS, and24 VRAM requirements before committing to it; use `gazebo` if that environment25 is not real and available.26- **MuJoCo:** prefer for lightweight, contact-rich manipulation that does not27 need ROS integration or photoreal rendering.28- A project may use different simulators for different concerns. Avoid forcing29 one tool to own both navigation validation and training-data generation.3031After selection, use `gazebo`, `isaac-sim`, or `mujoco` for tool mechanics.32Use `isaac-lab` for training workflows on top of Isaac Sim.3334## Define fidelity before building3536- Name what must transfer to hardware: interfaces, sensor behavior, dynamics,37 contacts, timing, or visual appearance.38- Match real sensor rate, noise, frame names, range, field of view, resolution,39 and timestamp source where those affect downstream decisions.40- Record known model gaps instead of treating simulation success as real-world41 evidence.42- If the simulator will back regression tests, control time step and randomness43 and judge behavior with tolerances rather than pixel-perfect coincidence.4445Read [sensor fidelity](SENSOR-FIDELITY.md) when adding sensors, preparing46sim-to-real work, or turning a scenario into a repeatable test. Use `data` when47the unresolved question is what training data to generate rather than how to48simulate it.4950## Done5152- The selected tool fits the available hardware and the project's main53 fidelity requirement.54- The simulation contract names which interfaces and physical properties must55 match reality.56- Every consumer uses a consistent clock and frame convention.57- Repeatability is designed where the simulation will be used as a test.58- Version and hardware requirements were checked in the current upstream59 [Gazebo](https://gazebosim.org/docs/),60 [Isaac Sim](https://docs.isaacsim.omniverse.nvidia.com/), or61 [MuJoCo](https://mujoco.readthedocs.io/) documentation.