IWB python package for robot simulation

• About
• Getting Started
• Running the Tests
• Usage
• Development
• Built Using
• TODO
• Contributing
• Authors
• Acknowledgments

There are the python package for iwb milling robot simulation, relying on ROS Noetic distribution. With module "robot" and other relevant modules, the package performs robot modeling, visualization, path planning, controlling, kinematics and dynamics calculation, robot simulation and computation. The robot simulation and computation functionalities are realized by IWB simulation package "iwbRBDL".

You can run script "demo.py" independently, which already applied some functionalies of robot simulation. Of course, with IWB_ROBOT class in module "robot", you can also apply those functionalies with your own program. You can check the tutorial Here and the API document Here

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

These instructions will get you a copy of the repository up and enable it to run with your system environment. See development for informations for developers.

The contents you need to walk through are:<br> (*: the same as IWB ROS package.)

• System Environment (*)
• Prerequisites
  • ROS Noetic(*)
  • Moveit!(*)
  • Python3
  • Pykdl
  • hrl-kdl
  • iwbRBDL
  • Others
• Installation

This instruction is the same as that in IWB ROS Package, you can skip it if you have already fulfilled all the requirements here.

The packages work with ROS Noetic, which is primarily targeted at:

• Ubuntu 20.04.4 LTS

Other systems are supported to varying degrees. (--- ROS Noetics)

Besides, if you are running everything with a virtual machine, the following configurations are recommended:

• 30 GB storage.
• TODO...

You can follow the Tutorial Here to setup a virtual machine.

The packages are based on ROS Noetic distribution. You may install the desktop version of ROS Noetic Here. Please choose Desktop-Full install, which is also recommanded on that page.

Every time you want to run the packages with terminal, use the following command to setup the ROS environment.

source /opt/ros/noetic/setup.bash

if you want to source a different version of ROS, this line would be:

source /opt/ros/ROS-DISTRO/setup.bash
# for example ROS 2 foxy:
source /opt/ros/foxy/setup.bash

You may have noticed that it's possible to avoid doing this every time opening a terminal by adding this line into your system "~/.bashrc". If you do so, please check if you append the line for the ROS version that of your use. (Remember this when transplant the project to a new ROS version, for example ROS 2 foxy)

Don't forget to try the turtle example provided in ROS tutorial to validate your installation.

Instead of running ROS from terminal, you can use ros python API to run ROS. The modules like "rospy" "roscore" or "rosnode" are quite helpful, which are installed during the installation of ROS. Our package will append the ROS python package directory to python module searching path for you (But now it's for Noetic. TODO: use system ROS-DISTRO to softcode this.).

You can skip this if you have followed the same one in IWB ROS packages' readme file.

The path & motion planning parts of the packages are realized by Moveit, which is the most widely used for robot manipulation. And it cooperates well with ROS.

You can install moveit ROS (binary) packages by execute the following line in your terminal:

# here for ros noetic
sudo apt install ros-noetic-moveit 

tips: You could always install missing ros packages using command line like:

sudo apt update
sudo apt install ros-ROS_DISTRO-ros_package_name
# for example
sudo apt install ros-noetic-ros-control

The Moveit Tutorial can help you to understand its planning functionalities. The Python interfaces of Moveit are explained Here, which is quite helpful. And the process to set up a robot configuration for Moveit is shown Here.

Besides, for transplanting project to ROS 2, you may need Moveit2 instead of Moveit.

Ubuntu 20.04 and other versions of Debian Linux ship with Python 3 pre-installed.

You can also install it manually. But if so, there will be issues with library path. The python you install manually should be in directory like "/usr/local/bin...", but the one that is original with OS is in "/usr/bin...".

optional: install pip for further package installing.

sudo apt update
sudo apt -y upgrade
sudo apt install -y python3-pip

Note that Pykdl should be installed with the installation of ROS. You can also install it using:

sudo apt update
sudo apt install python3-pykdl

Hrl-kdl provides us with lots of utils for KDL, which can help us to parse the robot URDF model directly and extract the kinematic chain with only several steps. Its Tutorial is Here, which is somewhat out of date. Our tutorial also provide you with a light introduction of it, check it Here.

For some reasons, it would be better to setup it from its repository clone. The following steps come from Here, note that the version should be "Python3 for Ubuntu 20.04 with ROS Noetic".

open a terminal and run:

cd
mkdir -p catkin_ws/src
cd ~/catkin_ws/src/
# clone the project repositroy (requires git tool)
git clone https://github.com/amir-yazdani/hrl-kdl.git

Checkout to the noetic-devel branch

cd hrl-kdl
git checkout noetic-devel

Install pykd_utils

cd pykdl_utils
# python3 setup.py build (maybe need sudo)
sudo python3 setup.py install

Install hrl_geom

cd ~/catkin_ws/src/hrl-kdl/hrl_geom/
python3 setup.py build
sudo python3 setup.py install

install urdf_parser and urdfdom-py

sudo apt-get install ros-noetic-urdf-parser-plugin
sudo apt-get install ros-noetic-urdfdom-py

Build the catkin workspace

cd ~/catkin_ws
# install catkin build tool package
sudo apt-get install python3-catkin-tools
# then build catkin
catkin build

Now the hrl-kdl has been succeessfully installed.

TODO

TODO

After fulfilling the prerequisites, you can now install the package to your device.

Open a terminal (ctrl+alt+T) and run the lines below to make new directory for the project.

cd
mkdir -p my_pkg
cd my_pkg

Then run git command to clone the packages into your new directory (You may need permission for that, e.g. ask the owner to add you as a maintainer.):

# Do not ignore the point "." at last!
git clone https://gitlab.lrz.de/00000000014A6C01/sa_py_pkg.git .

Now check the packages you cloned. If you have put the iwb_ros_package in another directory (not the "sa_ws" which the tutorial suggests), you need to set your custom directory before installation. Open /iwb_ros/script/iwb_ros_setting.sh and change the following content tnto yours.

# change this to your folder, no space on the left & right side of "="
path=PATH_OF_YOUR_FOLDER

The directory where you place the iwb_ros packages should be PATH_OF_YOUR_FOLDER/src

After this, now you can install the package using:

# Don't forget to change the directory into my_pkg
# You need sudo for that
sudo python3 setup.py install

It is recommended to check the installation path of your python package (library directory). You can check it with python terminal interaction. Open a new terminal and run (must ensure the work directory is not the one you placed our repository, otherwise the package in the repository may be imported instead of the one you installed):

python3

to start Python3 terminal interaction, then run the following in that:

# Now you should be able to import our package
import iwb_ros.setting
iwb_ros.setting.bash_strout("package_path")

The output should be similar like:

>>> iwb_ros.setting.bash_strout("package_path")
/usr/local/lib/python3.8/dist-packages/iwb_ros

which means your installation succeeds so far. You can use ctrl+d to quit the interaction programming of python3.

Since our package needs to run bash script to make use of the launch file of ROS packages, you need to change those scripts as executable. In terminal the command for that should be "chmod +x FILENAME", but here you can run a script to achieve it. Open a new terminal and run:

cd my_pkg
# use sudo to prevent taking current directory as python workspace
sudo bash system_permission.sh

Check the outputs of the terminal, which should always start with "/usr/local/lib/python3.8/dist-packages/iwb_ros/script/...". If they start with your current directory, open a new terminal and try to run bash command with absolute path using sudo.

Run it without error, then the package is ready for using.

You can run demo.py for checking. Note that since the documents needed are in /my_pkg, you should always run it from this directory. Because of the use of multi-thread functions, you should run it from IDEs, for which vscode is recommended.

Install vscode by running:

sudo snap install --classic code

Then open the directory /my_pkg in vscode. You also need to install the python extension for running the test. Please search the python extension in Extensions.

Then you can run the test script demo.py. The robot is visualized at first. Then the control node is launched, and the robot is set to the configuration for test. After that, the Jacobian matrix, mass matrix and eigenfrequencies are calculated and printed. Finally, the process is terminated clearly.

You can also use python terminal command to run it. Before doing that, you need to install the dependent libraries at first:

sudo pip3 install mat4py

then run the demo.py like:

cd my_pkg/
python3 demo.py

While running it from the terminal, there are many error reports of the velocity settings in pykdl, while the functions of the model are not affected. Besides, the environment of the terminal is dependent on the system setup, and debug is also not possible with the termianl. Therefore, this method is not recommanded.

TODO

The tree sturcture of the package is shown as follow:

TODO: tree structure

TODO API

• ROS Noeitcs - Robot operation system API
• Moveit - Motion planning tool
• Pykdl - Robot kinematics
• hrl-kdl - User interfaces of Pykdl
• iwbRBDL - IWB robot computation package

• @BBlab - Student in mechanical engineering
• And ... more contributers in the future.

• TODO: citing and inspiration mabe
• Thanks for your contribution to the project!
• ...