GRID Docs | General Robotics

In order to enable seamless integration of GRID into your current robotics stack, we provide ROS2 support with Isaac. This enables users to both publish data from the simulation as well as receive commands into the simulation to execute.

Publishing Simulation Data

In order to publish simulation data, we provide support for a list of message types the details of which are provided below:

Message Type	Configurable Parameters	Description
RGBImage	`msg_type: "sensor_msgs/Image"`, `target: "name_of_rgb_camera"`, `msg_name: "name_of_topic"`	RGB image frames
DepthImage	`msg_type: "sensor_msgs/Image"`, `target: "name_of_depth_camera"`, `msg_name: "name_of_topic"`	Depth image frames
BasePose	`msg_type: "geometry_msgs/Pose"`, `target: "name_of_robot"`, `body_name: "name_of_body"`, `msg_name: "name_of_topic"`	Pose of the robot’s base link
EndEffectorPose	`msg_type: "geometry_msgs/Pose"`, `target: "name_of_robot"`, `link_name: "name_of_link"`, `msg_name: "name_of_topic"`	Pose of the end-effector link (e.g. `panda_hand`)
JointState	`msg_type: "sensor_msgs/JointState"`, `target: "name_of_robot"`, `msg_name: "name_of_topic"`	Current joint states of the robot
ContactForces	`msg_type: "geometry_msgs/Wrench"`, `target: "name_of_contact_sensor"`, `msg_name: "name_of_topic"`	Contact force/torque readings
LidarScan	`msg_type: "sensor_msgs/PointCloud2"`, `target: "name_of_lidar"`, `msg_name: "name_of_topic"`	Point-cloud data from the LiDAR
ImuData	`msg_type: "sensor_msgs/msg/Imu"`, `target: "name_of_imu"`, `msg_name: "name_of_topic"`	IMU (accelerometer/gyroscope) measurements
GPSOdometry	`msg_type: "nav_msgs/msg/Odometry"` `target: "name_of_robot"` `body_name: "root"` `msg_name: "name_of_topic"` `info: {}`	Publishes full 6-DoF odometry (position, orientation, linear & angular velocity) of the robot’s root link, as from “GPS”.
Quaternion_Z_Relative	`msg_type: "geometry_msgs/msg/Quaternion"` `target: "name_of_height_sensor"` `msg_name: "name_of_topic"` `body_name: "root"` `info: { art_body: "robot" }`	Computes and publishes the Z-axis–only relative orientation quaternion between the robot and the ground plane under a raycaster sensor.
CameraInfo	`msg_type: "sensor_msgs/msg/CameraInfo"` `target: "name_of_rgb_camera"` `msg_name: "name_of_topic"` `info: {}`	Publishes camera calibration data (image size, distortion coefficients D, intrinsic K, rectification R, projection P) flattened.

An example of how to publish simulation data using env.yaml is as follows:

comms:
  - name_of_camera_msg:
      type: RGBImage
      config:
        target: camera_rgb_0 # name of the camera sensor
        msg_name: /robot/camera_rgb_0 # name of the topic to publish data

Subscribing to Commands

We also allow users to send commands to the robot in the simulation session. Here is the list of commands we support.

ROS2 Msg Type	Description
`sensor_msgs/JointState`	Command message to set robot joint positions
`geometry_msgs/Twist`	Command message to control base linear/angular velocity
`geometry_msgs/Pose`	Command message to set end-effector delta pose

For each command, users need to specify the name of the topic to which they are publishing the specific message and the target command name from the commands entity specified in the mdp.yaml. An example of how to use commands in env.yaml is as follows:

comms:
  - name_of_the_command_msg:
      type: Commands
      config:
        target: base_velocity # this is the name of the command 
        msg_name: /base_vel_cmd # this is the name of the topic to which the user is publishing

We are constantly updating our offering of supported message types, please let us know on our Discord Community or reach out to us on our email.

Get Started

Simulation

Deployment

FAQ

Support

ROS2 Communication

Publishing Simulation Data

Subscribing to Commands

​Publishing Simulation Data

​Subscribing to Commands

Publishing Simulation Data

Subscribing to Commands