Forward/inverse kinematics in Robot Operating System for the classes of the Artificial Intelligence bachelor studies.
If you don't have the ROS Noetic on your PC installed, you might follow the instructions under Installation/Using Docker to use Docker for that. Otherwise, you should set up ROS Noetic on your own and skip to the Installation/Manual installation.
-
You can just pull the Docker image:
docker pull mbed92/robotics_aiand create the container. To run create the container look into build_and_run.sh for more details (comment the building instruction - it's not needed anymore). -
Build and run the environment locally. Run the build_and_run.sh script with sudo priviliges (
sudo bash build_and_run.sh). After it finishes, it should create and enter into the container under the path /catkin_ws.
Since now everything should work smoothly and you can focus on your assignments. Also, you can open as many terminals as you need (for the classes two terminals are suggested - terminator is a very helpful tool for that) and attach each window to the created container using:
sudo docker exec -it robotics_ai bash
Now you have multiple terminals of the same environment.
Download that repositiory to the catkin_ws/src folder. Also, place there the software for the UR robotic manipulator.
sudo apt-get install -y python3-pip python3-catkin-tools python3-osrf-pycommon
cd /catkin_ws/src
git clone https://github.com/mbed92/robotics-ai-classes.git
git clone -b calibration_devel https://github.com/fmauch/universal_robot
sudo apt update -qq
rosdep update
rosdep install --from-paths src --ignore-src -y
source /catkin_ws/devel/setup.bash
In the course there are two assignments related to forward and inverse kinematics of a 6-DoF
robotic manipulator. Forward kinematics tasks are marked with a ### [TODO FK] ... comments,
while the inverse kinematics part is marked with ### [TODO IK] ... comments.
The task is to write the code that calculates the position and orientation of the green marker
with respect to the joint values (green box). As you move the robot with trackbars, the marker should
follow the tip of the robot. The repository contains code lines marked iwth a ### [TODO FK] ... comment.
Look into them and fill that places with your code:
- in move_robot/config/ur3_dh.yaml - define the UR3 robot with a DH parameters specified in link
- in move_robot/kinematics/KinematicsManager.py - define a subscriber for joint states and publisher for the marker position
- in move_robot/kinematics/KinematicsManager.py - implement the method that takes joint values and calculates the position and orientation of the end-effector
Run command:
roslaunch move_robot start_fk.launch
The task is to write the code that calculates the robot's pose (i.e. joint values) given the position and
orientation of the interaction marker (three axes with a red sphere handle). As you move the interaction
marker with a mouse, the robot should adapt its pose to follow the marker. Look into lines with ### [TODO IK] ... and
fill them with your code:
- in move_robot/kinematics/JointStatePublisher.py - implement your joint state publisher that publishes joint values from the inverse kinematics and updates the robot's pose
- in move_robot/kinematics/utilities.py - define a kinematic chain in pybotics
- in move_robot/kinematics/KinematicsManager.py - implement the inverse kinematics callback function that caluclate the joint values w. r. t. to the end-effector pose
- in move_robot/kinematics/KinematicsManager.py - write a publisher for calculated joint values
Inverse kinematics library - pybotics:
Run command:
roslaunch move_robot start_ik.launchContact me if you need any help with assignments: [email protected]