This repository uses NVIDIA's Isaac Sim for simulating the Adir robot.
- Docker
- NVIDIA Container Toolkit
- NVIDIA GPU with drivers installed
- NGC account
- ROS2 (Optional)
- Create an NGC API Key:
- Go to the NGC website.
- Log in with your NGC account.
- Navigate to the "Setup" page.
- Click on "Get API Key" and generate a new key.
- Login to Docker:
- Open a terminal.
- Run the following command to log in to the NVIDIA container registry:
- When prompted, enter your NGC username and the API key as the password.
Clone and Navigate to the project directory:
git clone https://github.com/AntRobotics-de/adir_simulation.git
cd adir_simulationRun the Docker script:
./run_simulation_docker.sh- Edit the run_simulation_local.sh script:
-
Open the script in a text editor:
nano run_simulation_local.sh
-
Update the ISAAC_SIM_HOME_DIR variable to the correct location of Isaac Sim on your system:
set_default ISAAC_SIM_HOME_DIR "/path/to/your/isaac-sim"Note: Replace
/path/to/your/isaac-simwith the actual path where Isaac Sim is installed on your system.
-
- Run the local script:
./run_simulation_local.sh
The following ROS2 topics are used in the Adir simulation:
/adir/front/livox/lidar(sensor_msgs/PointCloud2): Front LiDAR sensor data./adir/imu(sensor_msgs/Imu): IMU sensor data./adir/odom(nav_msgs/Odometry): Odometry data./adir/rear/livox/lidar(sensor_msgs/PointCloud2): Rear LiDAR sensor data./clock(rosgraph_msgs/Clock): Simulation clock./tf(tf2_msgs/TFMessage): Transform data./tf_static(tf2_msgs/TFMessage): Static transform data.
/cmd_vel(geometry_msgs/Twist): Command velocity for controlling the robot.