twolinkman

Two Link Manipulator control in ROS

Getting started

Prerequisites

• ROS Melodic with Gazebo-9 (Desktop-Full Install)
• Anaconda with Python 2.7 version

Installing

• Clone the repository to the local machine and change the working branch to 'ctc' OR download the .zip file of this branch
• Place the repository folder in /your/machine/catkin_ws/src
• Navigate into catkin_ws folder and open terminal.
• Type catkin_make and press enter.
• Make sure your .bashrc file (located in Home Folder) has these lines added in them: source yourpath/ros/melodic/setup.bash source yourpath/catkin_ws/devel/setup.bash
• After catkin has completed making the workspace, ROS should be able to locate your new rospackage twolinkman. To check it, type in terminal rospack find twolinkman. This should print out the location of the twolinkman_description folder. If the rospackage location is not displayed, please contact the collaborators.

Running the tests

• Open a terminal and type roslaunch twolinkman twolinkman_gazebo.launch. A Gazebo world should be loaded with a two link manipulator at the center. Minimize this terminal.
• Open another terminal and type roslaunch twolinkman twolinkman_control.launch. After several logging prints, the logging should stop at Started controllers: joint_state_controller, rear_joint_effort_controller, front_joint_effort_controller. Minimize this terminal.
• Open another terminal and type rostopic pub /twolinkman/front_joint_effort_controller/command std_msgs/Float64 "data: 1.0". This should move the front link of the two link manipulator in Gazebo world. Similarly try to change values of the topics by using rostopic pub and monitor angles using rostopic echo NOTE : If anything is logged in red color while running any of the above commands, there has been some error. Contact collaborators in such case with the log messages.

Python interface

• Navigate to Python directory in the repository and open a terminal.
• Type python roslauncher.py. The same world should be spawned in Gazebo as twolinkman_gazebo.launch file. Minimize the terminal
• Open a new terminal at same location and type python main.py. Logging as in twolinkman_control.launch file will be printed, followed by the Gazebo model which will move to desired joint angles and plotting a graph of current and desired joint angles.
• Feel free to look into main.py to change setpoints, gains, frequencies or complete controller. NOTE : While running the python files, module import errors may occur. Keep installing the missing modules using pip. If the problem persists, contact collaborators.

Making dataset

• If the setup has passed the tests mentioned in the above section, you can start collecting dataset for your manipulator model.
• Run the roslauncher.py and main.py scripts. The main.py scripts ends with a prompt requiring input of angle setpoints. Enter angles and the plotting will show the reponse along with performance measurements.
• The prompt pops up every 3 seconds.
• The dynamic reconfiguration of PID parameters can be done using rqt.
• Open a new terminal and type rqt.
• From the menu bar, go to Plugins > Configuration > Dynamic Reconfigure
• Select the controller which you want to tune and adjust PID parameters by typing in the textbox or scrolling the bars and press Enter.
• This enables live PID tuning during simulation of the manipulator.

NOTE: A more user friendly interface is under development.

Built with

• SolidWorks - 3D CAD model
• ROS Meloidc - ROS control node
• Gazebo 9 - Simulation environment
• Python 2.7 - Python interface for easy UI

Authors

• Aniket Sharma
• Bhavik Parmar
• Sagar Malik

Acknowledgements

Hearty thanks to Dr. S. N. Sharma for pointing out useful resources to refer to while working on this project. Deep gratitutde towards the mechanical team of the project for developing the experimental setup to test control schemes.