Docker file and docker-compose for ROS noetic supporting packages for the UR5e and Robotiq 2F-85 gripper. The compose file starts two containers: the one running ROS and another running noVNC to use the GUI-based tools rviz and Gazebo.
- Install either docker engine (Linux) or Docker Desktop.
- Clone this repo and navigate to the parent directory.
- Enable X11 forwarding
- Linux:
xhost +
- Windows: use VcXsrv
- Linux:
- Run
docker-compose up
.
- Make sure latest Nvidia drivers are installed
- Install nvidia-container-toolkit
- (Optional) Install nvidia-cuda-toolkit
roslaunch homestri_bringup a_bot_calibration.launch
roslaunch homestri_bringup a_bot_bringup.launch
roslaunch homestri_bringup a_bot_gazebo.launch
roslaunch realsense2_camera rs_camera.launch
- To ensure the camera is working, view the
/camera/color/image_raw topic
in RViz
- ensure
/camera_info
and/image
topics are properly remapped inaruco_ros/launch/double.launch
roslaunch aruco_ros double.launch
- Bringup the real robot
- Switch to velocity controller using
rqt_gui
rosrun rqt_gui rqt_gui
- In
rqt_gui
go toPlugins > Robot Tools > Controller Manager
- Stop
scaled_pos_joint_traj_controller
- Run
joint_group_vel_controller
- Launch joystick publisher
roslaunch homestri_moveit_servo joystick_teleop.launch
- Read and confirm joystick values and twist conversion
rostopic echo /joy
rostopic echo /servo_server/cmd_vel
- Launch
moveit_servo
roslaunch homestri_moveit_servo servo_server.launch
- Bringup the real robot (previous step)
- Start Aruco Tracking (previous step)
- Run manipulation_action_server
rosrun homestri_manipulation manipulation_action_server.py
- If you are grasping with an offset from a tag, use a static transform publisher to consider the offset (This will be changed later)
- Launch Explainable Behavior Tree UI
rosrun rqt_gui rqt_gui
- In
rqt_gui
go toPlugins > Actions > Explain BT Plugin
- Execute behavior tree
roslaunch homestri_behavior_trees my_bt.launch