Sensors | Isaac Lab

These pages are the source of truth for the isaaclab-using-sensors-actuators agent skill (skills/user/use-sensors-actuators/). When sensor behavior changes, keep the API documentation, maintained demos, these concept pages, and the skill synchronized.

Sensors turn simulation state or rendered scene data into batched measurements. Every Isaac Lab sensor derives from ~isaaclab.sensors.SensorBase and follows the same lifecycle:

prim_path selects the prim or prims measured in every cloned environment.
update_period sets the sampling period in simulated seconds. A value of 0.0 samples every simulation step.
~isaaclab.sensors.SensorBase.update advances the sensor clock. Sensor data is evaluated lazily when the isaaclab.sensors.SensorBase.data property is read, unless recomputation or debug visualization is requested.
~isaaclab.sensors.SensorBase.reset clears per-environment timestamps and internal state.

With eager sensor updates (scene.cfg.lazy_sensor_update=False), scene.update(dt) refreshes sensor data before returning. It advances sensor clocks in scene order, updating ordinary sensors individually and collecting sensors that support batch updates. After the loop, it refreshes the collected sensors’ pending buffers together. Each sensor retains its own update period and reset state. Camera sensors use this mechanism to batch captures by renderer.

Sensor data is exposed through ~isaaclab.utils.warp.ProxyArray buffers, including camera outputs. Use the torch property for a cached zero-copy Torch view or warp for the underlying Warp array.

Choose a sensor

Sensor	Measurement	Typical use
camera	Renderer-produced color, depth, normals, motion, and segmentation images	Vision observations and synthetic data
ray_caster	First collision locations for configurable ray patterns	Terrain sensing, lidar-style queries, and ray-cast cameras
contact_sensor	Net and filtered contact forces, contact points, friction, and contact timing	Locomotion, grasping, and collision events
frame_transformer	Relative and world poses for configured frames	End-effector, foot, and object tracking
imu	Angular velocity and proper linear acceleration	Inertial observations
pva	Ground-truth pose, velocity, coordinate acceleration, and projected gravity	State estimation and privileged observations
joint_wrench_sensor	Incoming joint reaction force and torque	Force/torque sensing and contact-rich control