ROS MoveIt Robotic Arm

in robot •  3 months ago


This is going to be a series of articles about  making a robotic arm controlled with ROS(Robotic Operating System) and  MoveIt. If you are into robotics ROS is a great framework that will help  you to build better robots faster. It allows you to re-utilize packages  for different robot systems(computer vision, kinematics, path planning,  SLAM, etc) that other people created. ROS is great, but unfortunately  it does have a steep learning curve. So this tutorial is  step-by-step guide to making your first robotic arm with motion  planning, aimed at upper-beginner and intermediate levels. I assume you  have no knowledge of ROS, but do have some basic knowledge Linux and  Python, Arduino programming skills.

 Step 1: Prepare Your Environment

 My environment for development is Ubuntu 16.04 LTS running in a  Virtual Box(Windows 10 host machine). If you want to run Ubuntu on VM  it's fine, I can confirm ROS Kinetic works, albeit with some  quirks(notably 3D acceleration and USB support). If you have regular  Ubuntu installation, it would work better. Note: You can install  ROS on Raspberry Pi, but unfortunately it's not powerful enough to run  the simulations in Rviz, let alone Gazebo. 

1) Install ROS Kinetic. Follow this guide on how to install ROS Kinetic on Ubuntu 16.04. 

2) Install MoveIt. MoveIt! is a package for ROS for mobile manipulation, i.e. robotic arms. 

3) Check if everything works. Run 

roslaunch panda_moveit_config demo.launch rviz_tutorial:=true

if you want, go through moveit basic tutorial
4) You'll need to install urdf_tutorial package as well, we'll use it to visualize our work in progress.

sudo apt-get install ros-kinetic-urdf-tutorial

 

Step 2: Create Urdf File

 If you followed MoveIt quickstart tutorial you  should have created catkin workspace in your home directory(the name of  the workspace might be different, default from moveit tutorial is  ws_moveit). Go to ~ws_moveit/src and clone my github repo there.

https://github.com/AIWintermuteAI/ros-moveit-arm.git

If  you are using the same robotic arm as I do or just want to practice you  can leave the urdf file unchanged, otherwise you'll need to change it  to correspond the arm model you have.

Here's a brief  explanation of urdf file, looking at my example and the robot model in  rviz will help you to understand it better and apply for writing your  own urdf file. 

 <link> links are the parts of the robot, like a wrist or a platform. They have following properties:  

<visual> used for visual representation in rviz and gazebo. Uses high-quality meshes    

<collision> used for collision checking during path planning.  Recommended to use low-quality meshes, but for tutorial's sake I'm using  the same as for visual.  

<joint> joints in urdf are  connections between links which describe how the are positioned in  relation to each other and how they move. They have following  properties: 

<axis> the rotation direction of the joint(along x, y or z axis)  

<limit effort> minimum and maximum degree of the joint in  radians. Only applicable to "revolute" type joints("continuous" joints  have no limits, since they, well, rotate continuously) 

<origin> the position of the joint in relation to the origin 

<parent link> <child link> describes which two links are connected with the jointIf  you are using the same robotic arm as I do, you'll need to copy 3D  meshes of the arm to urdf_tutorial package folder. Do it with this  command

sudo cp <names of hte files to copy> /opt/ros/kinetic/share/urdf_tutorial/meshes/

If  you're using your own meshes make sure they are centered around the  origin(xyz 000 in your favorite 3D modeling software)  before you export  them to binary(!) stl format. Now you can launch rviz to visualize the results with this command 

roslaunch urdf_tutorial display.launch model:=arm.urdf

(launch it from the folder with the urdf file you created just now)Check  the joints and the links, make sure they are rotating and positioned  exactly how they are supposed to. If everything is okay, move on to the  next step. 

 

Step 3: Create Robot Configuration With MoveIt Setup Assistant

 Now when your urdf file is ready it's time to convert it to robot configuration package. Launch MoveIt Setup Assistant with the following command 

roslaunch moveit_setup_assistant setup_assistant.launch

The process is very well described in MoveIt official tutorial, which you can find here
After you created the configuration package, go to your catkin workspace folder and execute 

catkin make

to build the package. Now you can launch it with the following command 

roslaunch my_arm_xacro demo.launch rviz_tutorial:=true

where you will need to change "my_arm_xacro" to the name of your package. 

 

Step 4: Done!

 

If you successfully launched the package  you'll see your robot model with interactive markers. You'll be able to  control it in simulation, using the markers and then press "Plan and  execute" for the kinematic solver to find hte right path to the desired  state.That was the step one, we created a model of our robotic  arm and can harvest the power of ROS to calculate and execute motion in  the simulation! Next step - do it with the real robot, for that we'll  need to write simple robot controller...

Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!
Sort Order:  

Nice stuff. ROS is on my list of things to learn anyway...wanted to try this for one of my robots. Can't wait to read more about it.

·

Your InMoov robot is awesome! How much time did it take you to build it?
Here's the second part of the article
https://steemit.com/robotics/@wintermuteai/ros-moveit-robotic-arm-part-2-robot-controller

·
·

I only did it in my sparetime so it took about 3 years.

Congratulations This post has been upvoted by SteemMakers. We are a community-based project that aims to support makers and DIYers on the blockchain in every way possible.

Join our Discord Channel to connect with us and nominate your own or somebody else's posts in our review channel.

Help us to reward you for making it ! Join our voting trail or delegate steem power to the community account.

Your post is also presented on the community website www.steemmakers.com where you can find other selected content.

If you like our work, please consider upvoting this comment to support the growth of our community. Thank you.