Project repository for building extensible robotic beamline scientists at NSLS-II.
The majority of the contents in this repository are ROS2 packages with associated continer image manifests. Each manifest in the docker directory is a container image that can be used to run a specific application in the system.
- pdf_beamtime: ROS2 package for controlling the UR3e robot arm and HandE gripper at the PDF beamline.
- pdf_beamtime_interfaces: ROS2 package for defining the interfaces used in the pdf_beamtime package.
- hello_moveit: ROS2 package for demonstrating simple actions using the MoveIt library with the UR3e robot arm.
- hello_moveit_interfaces: ROS2 package for defining the interfaces used in the hello_moveit package.
- bluesky_ros: Python module for integrating Bluesky and ROS2.
- gripper_service: ROS2 package for controlling the HandE gripper.
- ur3e_hande_robot_description: ROS2 package for defining the UR3e robot arm with the HandE gripper.
- ur3e_hande_moveit_config: ROS2 package for configuring MoveIt for the UR3e robot arm with the HandE gripper.
- aruco_pose: ROS2 package for detecting ArUco markers and calculating their pose.
- ros2.repos: ROS2 workspace file for downloading the required external ROS2 dependencies.
We use Podman throughout this work, but have named the container images with Docker in mind.
- erobs-common-img: Common container image for running the majority of applications herein, including: UR robot driver, gripper service, MoveIt service, and the pdf_beamtime_server (primary
Action
server). - ursim: Container image for running a simulated UR3e robot arm with a teach pendant.
- ur-driver: Container image for running the UR3e robot arm ROS2 driver.
- ur-moveit: Container image for running MoveIt with the UR3e robot arm.
- ur-example: Container image for running a simple action with the UR3e robot arm.
- ur-hande-draft: Deprecated container image for running the UR3e robot arm with the HandE gripper.
- erobs-hello-moveit: Container image for running a simple action with the UR3e robot arm.
- bsui: Container image for running the Bluesky User Interface with mounts at NSLS-II.
- azure-kinect: Container image for running the Azure Kinect ROS2 driver.
Demonstrations using a combination of the MoveIt tutorials and some UR specific tools, to show how to make simple actions that can deploy MoveIt using the MoveGroupInterface.
Ongoing developments of integrating ROS2 and Bluesky. Currently targeted towards integrating Ophyd Objects as ROS2 Action Clients.
The complete application uses a 1-node-per-container model. The containers are currently orchestrated by bash scripts detailed in the READMEs of each container image. Specifically, the full application is detailed in erobs-common-img.
In order to run the ur-example
with Docker, follow this procedure:
- Create the required images.
cd docker
docker build -t ursim:latest ./ursim
docker build -t ur-driver:latest ./ur-driver
docker build -t ur-example:latest ./ur-example
- Start the UR Simulator. In a new terminal, run
docker compose up ursim
-
Open VNC client at
localhost:5900
. -
Turn on and start the robot.
-
Go to the
Move
tab and click theHome
button. -
Press and hold the
Move robot to: New position
button to move the robot into position. PressContinue
. -
Verify the joint position is
[0, -90, 0, -90, 0, 0]
degrees. -
Note: setting initial position is requried for the
ur-example
to start, as specified in thetest_goal_publisher_config.yaml
file in the official Unviersal_Robots_ROS2_Driver repo. -
Start the ur-driver. In a new terminal, run
docker compose up urdriver
Now, go back to the VNC client. In the Program
tab, start the program.
- Run the ur-example. In a new terminal, run
docker compose up urexample
The in Program/Graphics
tab, the robot should be moving between four poses every 6 seconds.
VSCode ROS2 Workspace Template Borrowed from @althack.
This template will get you set up using ROS2 with VSCode as your IDE. And help ensure consistent development across the project.
See how she develops with vscode and ros2 for a more in-depth look on how to use this workspace.
ROS2-approved formatters are included in the IDE.
- c++ uncrustify; config from
ament_uncrustify
- python autopep8; vscode settings consistent with the style guide
pdf_beamtime is a work-in-progress package aiming to deploy the UR3e robot arm + HandE gripper at the PDF beamline. This package depends on pdf_beamtime_interfaces. Follow the link below for information on the package and for the commands to call the servers implemented in the package.