GRID Docs | General Robotics

GRID currently supports the following sensors with Isaac Sim.

Camera

This is a camera sensor that can be of type RGB or Depth.

Configuration

The configuration for the camera sensor can be set up in the configuration file. The following parameters are available:

Parameter	Description
Height	The height of the captured image
Width	The width of the captured image
Offset	The offset of the camera with respect to the robot’s frame of reference
Focal Length	The focal length of the camera
Focal Distance	The focal distance of the camera
Horizontal Aperture	The horizontal aperture of the camera
Clipping Range	The clipping range of the camera

These parameters can be configured by adding the sensor in the scene entity inside env.yaml in the following manner:

- camera_sensor_name:
    type: Sensor
    name: CameraRGB
    config:
        prim_path: "/World/envs/env_.*/tabletop_rgb_camera"
        height: 256
        width: 256
        offset:
            pos: [1.5, 0.0, 0.7]
            rot: [0.0, -0.3, 0.0, 1.0]
            convention: "world"
        spawn:
            focal_length: 24.0
            focus_distance: 400.0
            horizontal_aperture: 20.955
            clipping_range: [0.1, 100000]

Contact Sensor

This is a simple contact sensor that works as a ray caster simulating surface pressure by detecting contact points with precision.

Configuration

The configuration for the height sensor can be set up in the configuration file. The following parameters are available:

Parameter	Description
History length	The length of history to consider while computing sensor values

These parameters can be configured by adding the sensor in the scene entity inside env.yaml in the following manner:

- contact_sensor_name:
    type: Sensor
    name: ContactSensor
    config:
        prim_path: "/World/envs/env_.*/robot/.*"
        history_length: 3

Note that you need to make sure activate_contact_sensors flag is set to true for your robot.

Height Sensor

This is a simple height sensor that works as a ray caster to obtain the height values of points in a specified pattern.

Configuration

The configuration for the height sensor can be set up in the configuration file. The following parameters are available:

Parameter	Description
Offset	The offset of the height sensor with respect to robot’s frame of reference
Pattern Resolution	The resolution of the grid pattern around the sensor
Pattern Size	The start and end of the grid pattern

These parameters can be configured by adding the sensor in the scene entity inside env.yaml in the following manner:

- height_sensor_name:
    type: Sensor
    name: HeightSensor
    config:
        prim_path: "/World/envs/env_.*/robot/body"
        offset:
            pos: [0.3, 0.0, 0.22]
            rot: [0.0, 0.0, 0.0, 1.0]
        pattern_cfg:
            resolution: 0.1
            size: [1.6, 1.0]

LiDAR

This is a LiDAR sensor that works as a ray caster to obtain a point cloud around the sensor.

Configuration

The configuration for the height sensor can be set up in the configuration file. The following parameters are available:

Parameter	Description
Offset	The offset of the height sensor with respect to the robot’s frame of reference
Pattern Size	The start and end of the grid pattern
Max Distance	The maximum sensing distance of the height sensor
Vertical FOV	The vertical field of view of the sensor
Horizontal FOV	The horizontal field of view of the sensor
Horizontal Resolution	The horizontal resolution of the sensor
Channels	The number of channels used by the sensor

These parameters can be configured by adding the sensor in the scene entity inside env.yaml in the following manner:

- lidar_sensor_name:
    type: Sensor
    name: LiDAR
    config:
            offset:
                pos: [0.0, 0.0, 0.0] # add init (x, y, z) position
                rot: [0.0, 0.0, 0.0, 0.0] # add init (x, y, z, w) quaternion
            pattern_cfg:
                channels: 16
                vertical_fov_range: [-15.0, 15.0]
                horizontal_fov_range: [-180.0, 180.0]
                horizontal_res: 0.2
            max_distance: 100

IMU

The Inertial Measurement Unit (IMU) provides orientation and acceleration data by measuring the robot’s linear accelerations and angular velocities in its own body frame.

Configuration

The configuration for the IMU sensor can be set up in the configuration file. The following parameters are available:

Parameter	Description
Offset	Pose of the IMU relative to the robot’s frame of reference
Gravity Bias	The constant gravity vector to subtract from raw acceleration readings, as a 3-tuple `(x, y, z)` (m/s²)

These parameters can be configured by adding the sensor in the scene entity inside env.yaml in the following manner:

- imu_sensor_name:
    type: Sensor
    name: IMU
    config:
            prim_path: "/World/envs/env_.*/robot/body"
            offset:
                pos: [0.0, 0.0, 0.0] # add init (x, y, z) position
                rot: [0.0, 0.0, 0.0, 0.0] # add init (x, y, z, w) quaternion
            gravity_bias: [0.0, 0.0, 9.81]

Rectified Camera

This is a rectified camera sensor that applies intrinsic and extrinsic calibration to produce undistorted images (e.g., for stereo rigs or calibrated monocular setups).

Configuration

The following parameters are available on the RectifiedCameraCfg:

Parameter	Description
Width	The width of the captured image
Height	The height of the captured image
Offset	The offset of the camera with respect to the robot’s frame of reference
D	Distortion coefficients `[k1, k2, p1, p2, (k3)]`.
K	Intrinsic matrix, flattened `[fx, 0, cx, 0, fy, cy, 0, 0, 1]`.
R	Rectification (rotation) matrix, flattened 3×3.
P	Projection matrix, flattened 3×4 `[fx′, 0, cx′, Tx, 0, fy′, cy′, Ty, 0, 0, 1, 0]`.
Clipping Range	The clipping range of the camera
Focal Length	The focal length of the camera
Focal Distance	The focal distance of the camera
F Stop	Aperture setting for spawn config (`0.0` means wide open).
Projection Type	Type of image projection — one of “pinhole”`or`”fisheye”.
Lock Camera	Whether to lock camera’s transform or not.

Example

Add a rectified camera to your env.yaml like so:

- rectified_camera:
    type: Sensor
    name: RectifiedCamera
    config:
          prim_path: "/World/envs/env_*/stereo_camera"
          width: 640
          height: 480
          offset:
            pos: [0.0, 0.0, 1.2]
            rot: [0.0, 0.0, 0.0, 1.0]
          D: [0.1, -0.05, 0.001, 0.0005, 0.0]
          K: [600.0, 0.0, 320.0,
               0.0, 600.0, 240.0,
               0.0,   0.0,   1.0]
          R: [1.0, 0.0, 0.0,
              0.0, 1.0, 0.0,
              0.0, 0.0, 1.0]
          P: [600.0, 0.0, 320.0, 0.0,
              0.0, 600.0, 240.0, 0.0,
              0.0,   0.0,   1.0, 0.0]
          clipping_range: [0.1, 100000]
          focal_length: 24.0
          focus_distance: 500.0
          f_stop: 2.8
          projection_type: "pinhole"
          lock_camera: true

Get Started

Simulation

Deployment

FAQ

Support

Sensors

Camera

Configuration

Contact Sensor

Configuration

Height Sensor

Configuration

LiDAR

Configuration

IMU

Configuration

Rectified Camera

Configuration

Example

​Camera

​Configuration

​Contact Sensor

​Configuration

​Height Sensor

​Configuration

​LiDAR

​Configuration

​IMU

​Configuration

​Rectified Camera

​Configuration

​Example

Camera

Configuration

Contact Sensor

Configuration

Height Sensor

Configuration

LiDAR

Configuration

IMU

Configuration

Rectified Camera

Configuration

Example