lihuang3 / Ur5_ros Gazebo

Implementation of UR5 pick and place in ROS-Gazebo with a USB cam and vacuum grippers.

This repository demonstrates UR5 pick-and-place in ROS and Gazebo. The UR5 uses a USB cam to detect a red box on a conveyor (ur5_vision.py), and publish its position. UR5 plans its motion (ur5_mp.py) to follow the box. Once the end-effector gets close enough to the box, it approaches the box with vacuum grippers turning on (ur5_gripper.py). Since the vacuum gripper only provides limited force, we placed multiple grippers in order to lift the object.

• Video demos: Simulation video Hardware video

• Hardware implementation in UR3: ur3_ROS-hardware

• Update: based on feebacks from the community, we have made several key changes to this repository on 09/16/2018. Please update your code in case you have trouble reproducing the results.

• How to cite this repository:

    Huang, L., Zhao, H., Implementation of UR5 pick and place in ROS-Gazebo with a USB cam and vacuum grippers, (2018), GitHub repository, https://github.com/lihuang3/ur5_ROS-Gazebo.git
  

  or BibTex

    @misc{Huang2018,
      author = {Huang, L., Zhao, H.},
      title = {Implementation of UR5 pick and place in ROS-Gazebo with a USB cam and vacuum grippers},
      year = {2018},
      publisher = {GitHub},
      journal = {GitHub repository},
      howpublished = {\url{https://github.com/lihuang3/ur5_ROS-Gazebo.git}}
    }
  

How to use this repository

• This project was tested in Ubuntu 16.04 with ROS kinetic.
• Make sure you have installed Python2.7 and some useful libraries/packages, such as Numpy, cv2, etc.
• Install ROS kinetic, Gazebo, universal robot, Moveit, RViz.
• Assuming your universal robot workspace is named as ur_ws, download the repository to ur_ws/src/

  $ cd ur_ws/src
  $ git clone https://github.com/lihuang3/ur5_ROS-Gazebo.git
  

• Under ur_ws/src, there are two folders: one is the official universal_robot, and the other is ur5_ROS-Gazebo. Open file ur5_joint_limited_robot.urdf.xacro under ur_ws/src/universal_robot/ur_description/urdf/, and make the following change to the joint limit:

    shoulder_pan_lower_limit="${-2*pi}" shoulder_pan_upper_limit="${2*pi}"
  

• In the same directory, make a copy of common.gazebo.xacro and ur5.urdf.xacro in case of any malfunction. These two default files do not include camera and vacuum gripper modules. So we would replace these two files with customized files. Under directory ur_ws/src/ur5_ROS-Gazebo/src/ur_description/, copy common.gazebo.xacro and ur5.urdf.xacro to ur_ws/src/universal_robot/ur_description/urdf/.
• Build the code under directory ur_ws/,

  $ catkin_make
  $ source devel/setup.bash  
  

• Run the code with ROS and Gazebo

  $ roslaunch ur5_notebook initialize.launch 
  

• Things to work on: (1) vacuum grippers only provide limited force for lifting, so we had to use so many of them in order to pick up a light box. If you have any suggestions, please let us know. (2) UR5 motion planning is not in realtime, and hence you can ovserve a non-smooth motion of the end-effect in the camera view.

0. References

• GitHub: utecrobotics/ur5 testing ur5 motion
• Very useful ROS blog ROS探索
• ROS下如何使用moveit驱动UR5机械臂
• ROS UR industrial
• ROS modern driver
• ur_hardware_interface.cpp Replace this cpp if there is hardware interface error during catkin make

######## Warning! The rest of this README is still under construction ###########

1. Universal Robot 5 Installation

Official installation tutorial

The following command lines are for ur5 installation in ros-kinetic

mkdir -p ur5_ws/src
cd ur5_ws/src

# retrieve the sources
git clone -b kinetic-devel https://github.com/ros-industrial/universal_robot.git

cd ~/ur5_ws

# checking dependencies
rosdep install --from-paths src --ignore-src --rosdistro kinetic

# buildin,
catkin_make

# if there is any error, try
# pip uninstall em
# pip install empy

# source this workspace (careful when also sourcing others)
cd ~/ur5_ws
source devel/setup.bash

To run UR5 in Gazebo and rviz (source devel/setup.bash)

roslaunch ur_gazebo ur5.launch limited:=true

roslaunch ur5_moveit_config ur5_moveit_planning_execution.launch sim:=true limited:=true

roslaunch ur5_moveit_config moveit_rviz.launch config:=true

To test UR5 motion, run testmotion.py Look into this link for straight line motion

2. Moveit

Official tutorial

2.0 Install Moveit

sudo apt-get install ros-kinetic-moveit

launch teh Moveit Setup Assistant

roslaunch moveit_setup_assistant setup_assistant.launch

1. Click on the "Create New MoveIt Configuration Package" button,click the "Browse" button, select the xacro file you created in the Previous Chapter, and click on the "Load Files" button.
PATH: /src/universal_robot/ur_description/urdf/ur5.urdf.xacro

2. Go to the "Self-Collisions" tab, and click on the "Regenerate Collision Matrix" button.

3. move to the "Virtual Joints" tab. Here, you will define a virtual joint for the base of the robot. Click the "Add Virtual Joint" button, and set the name of this joint to FixedBase, and the parent to world.

4. open the "Planning Groups" tab and click the "Add Group" button. Now, you will create a new group called manipulator, which uses the KDLKinematicsPlugin.

3. Move Group Python InterFace Tutorial[`Official tutorial`](http://docs.ros.org/indigo/api/moveit_tutorials/html/doc/pr2_tutorials/planning/scripts/doc/move_group_python_interface_tutorial.html)

2.1 Use Moveit Interface

Move Group Interface Tutorial

2.2 Use Moveit in Python

Cartesian Path Planning

Using the ur5 with the MoveIt Motion Planning Framework for quick motion planning. Install the package from package management, and run the MoveIt! planning demo:

$ sudo apt-get install ros-kinetic-ur5-moveit-config

$ roslaunch ur5_moveit_config demo.launch

Our goal is to move the universal robot (ur5) end effector moving in straight line (Cartesian path) with Moveit-Python interface.

References
[1] CMobley7 commented on ros-planning/moveit_commander
[2] homesick-nick UR5CubicInterpolation
[3] Move Group Python Interface Tutorial
[4] ur_modern_driver

3 USB Camera Installation in ROS

Reference link

To list all video devices picked up by the kernel

$ ls -ltrh /dev/video*

$ cd ur5_ws/src

$ git clone https://github.com/bosch-ros-pkg/usb_cam.git

$ cd ..

$ catkin_make

$ source devel/setup.bash

$ roscd usb_cam

# run `roscore` in a new terminal
# Make sure a usb cam is connected

To connect external cam. Locate the usb_cam-test.launch file in folder

cd ~/ur5_ws/src/usb_cam/launch

Change

<param name="video_device" value="/dev/video0" />

to

<param name="video_device" value="/dev/video1" />

From

cd ~/catkin-ws/src/usb_cam/launch run

roslaunch usb_cam-test.launch

If this works, quit the test program, open rviz

rosrun rviz rviz

run the following command in ur5_ws folder (source devel/setup.bash)

rosrun usb_cam usb_cam_node

4. Revolute-Revolute Manipulator Robot

"RRBot, or ''Revolute-Revolute Manipulator Robot'', is a simple 3-linkage, 2-joint arm that we will use to demonstrate various features of Gazebo and URDFs. It essentially a double inverted pendulum and demonstrates some fun control concepts within a simulator."

To get RRBot, clone the gazebo_ros_demos Github repo into the /src folder of your catkin workspace and rebuild your workspace:

cd ~/catkin_ws/src/
git clone https://github.com/ros-simulation/gazebo_ros_demos.git
cd ..
catkin_make
source devel/setup.bash

Quick start

Rviz:

roslaunch rrbot_description rrbot_rviz.launch

Gazebo:

roslaunch rrbot_gazebo rrbot_world.launch

ROS Control:

roslaunch rrbot_control rrbot_control.launch

Example of Moving Joints:

rostopic pub /rrbot/joint2_position_controller/command std_msgs/Float64 "data: -0.9"

5. Using Gazebo Camera Plugins

• Tutorial: Using a URDF in Gazebo prerequisite for Gazebo plugins

• Tutorial: Using Gazebo plugins with ROS learning to use Gazebo plugins

• Viewer for ROS image topics

• Camera implementation with IRIS drone

• ur_description wiki To view and manipulate the arm models in rviz, install the package from package management and launch the following:

roslaunch ur_description ur5_upload.launch
roslaunch ur_description test.launch

You probably need to install urdf_tutorial:

cd ~/catkin_ws/src/
git clone https://github.com/ros/urdf_tutorial
cd ..
catkin_make
source devel/setup.bash

In rviz, the first task is to choose the frame of reference for the visualization. In left panel, Global Options/Fixed Frame, choose a proper frame (e. g. world)

Next, we want to view the 3D model of the robot. To accomplish this, we will insert an instance of the robot model plugin To add the robot model to the rviz scene, click the “Add” button and choose RobotModel

To test UR5 USB cam, run testvision.py

5. gazebo world Launch

• Defining Joints in Solidworks to URDF Exporter for a Conveyor Belt

• Tutorial: simulator in gazebo

launch gazebo files :

roslaunch ur5_notebook initialize.launch

change object pose and twist with command line:

rosservice call /gazebo/set_model_state '{model_state: { model_name: red_box, pose: { position: { x: 0, y: 0 ,z: 1 }, orientation: {x: 0, y: 0, z: 0, w: 0 } }, twist: { linear: {x: 0.1 , y: 0 ,z: 0 } , angular: { x: 0.0 , y: 0 , z: 0.0 } } , reference_frame: world } }'

6. openCV

• Detect red circles in an image using OpenCV

• OpenCV moment

• OpenCV moment

• cv2.contourArea() examples

7. ariac

• ARIAC 2017: INSTALL 4.2 4.3