SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsisaaclab-using-sensors-actuators

isaaclab-using-sensors-actuators

General & other

View repository

Updated 10-15 22:38

About this skill

Using Sensors And Actuators

When To Use

Use this skill when a user wants to add, configure, or debug sensors or actuator models in an Isaac Lab task.

Do not use this skill as a complete sensor or actuator catalog. Point to the API docs and maintained examples for available classes and parameters.

Workflow

  1. Identify the data needed by the task: contacts, ray hits, camera images, IMU data, joint wrench data, actuator dynamics, or low-dimensional robot state.
  2. Read the relevant sensor or actuator docs before editing configs.
  3. Check whether the sensor requires a simulation app, renderer, backend-specific support, or special update period.
  4. Add sensor or actuator configs to the environment scene or asset config using existing task examples as templates.
  5. Register the sensor in the scene setup for direct workflows, or in the scene config for manager-based workflows.
  6. Add observation terms only after confirming the sensor data shape and device.
  7. For an existing task, inspect its scene config and existing MDP terms before adding code. Make the smallest task-local config change and create a shared MDP module only when the term is genuinely reusable.
  8. For contact-heavy tasks, verify body name patterns, history length, and whether track_air_time=True is required. Reuse maintained terms such as mdp.feet_air_time when they match the task.
  9. For ray-cast terrain perception, verify mesh paths and terrain import setup.
  10. For camera-based RL, start with small environment counts and confirm renderer memory behavior.
  11. For multi-backend contact, ray, frame, IMU, PVA, or joint-wrench sensors, use the common config from isaaclab.sensors; the active physics backend selects the implementation automatically. Use backend-specific fields or configs only when the current API or maintained source example requires them.
  12. For actuator changes, compare default joint names, limits, stiffness, damping, effort limits, and backend differences.

Validation

Use this checklist:

  1. Instantiate the environment with a small number of environments.
  2. Confirm the sensor or actuator config binds to the intended prims or joints.
  3. Step the simulation and inspect data shape, dtype, device, and update timing.
  4. Confirm observations include the intended data and no stale values.
  5. Run a short rollout or training smoke test after shape validation.

For skill changes, run:

uv run --no-project python tools/skills/cli.py check

Maintenance

Keep this skill synchronized with actuator docs under docs/source/concepts/, sensor docs under docs/source/concepts/sensors/, sensor tutorials under docs/source/how-to/, API docs under docs/source/api/, and maintained task examples under source/isaaclab_tasks/isaaclab_tasks/. If sensor support changes for PhysX, Newton, or renderers, update the docs or source examples first.

References