The goal of this lab is to teach you about self-contained calibration of the robots and refresh your knowledge about numerical optimization. The presentation can be found here and additional scripts for the Levenberg-Marquardt algorithm here.

Calibration papers from our group:

J. Rozlivek, L. Rustler, K. Stepanova and M. Hoffmann, "Multisensorial robot calibration framework and toolbox," 2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids), 2021, pp. 459-466.

K. Stepanova, J. Rozlivek, F. Puciow, P. Krsek, T. Pajdla, M. Hoffmann, Automatic self-contained calibration of an industrial dual-arm robot with cameras using self-contact, planar constraints, and self-observation, Robotics and Computer-Integrated Manufacturing, Volume 73, 2022.

K. Stepanova, T. Pajdla and M. Hoffmann, "Robot Self-Calibration Using Multiple Kinematic Chains—A Simulation Study on the iCub Humanoid Robot," in IEEE Robotics and Automation Letters, vol. 4, no. 2, pp. 1900-1907, April 2019.

A. Roncone, M. Hoffmann, U. Pattacini and G. Metta, "Automatic kinematic chain calibration using artificial skin: Self-touch in the iCub humanoid robot," 2014 IEEE International Conference on Robotics and Automation (ICRA), 2014, pp. 2305-2312.

L. Rustler, B. Potocna, M. Polic, K. Stepanova and M. Hoffmann, "Spatial calibration of whole-body artificial skin on a humanoid robot: comparing self-contact, 3D reconstruction, and CAD-based calibration," 2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids), 2021, pp. 445-452.

This lab is an introduction to a Matlab toolbox which provides a solution to the multi-chain calibration of a general robot by combining multiple calibration approaches.

You will need a Matlab (version at least 2020b). You can install Matlab from CTU download page. You also need to install the Optimization toolbox and Symbolic toolbox (all is included in the campus-wide license). If you do not want to install anything, you can use Matlab online. However, you still need to activate your Mathworks account as described on CTU download page.

• If you get download error in the middle of installation, you can download full ISO
  • in the download page, select Get ISOs instead of Install products in I WANT TO drop box
• The default path for installation (/usr/local/MATLAB) belongs to root, so you will probably have to run the installer with sudo. However, several errors can occur:
  • licence problem → do not forget to select correct user when prompted during the installation—insert your username (can be obtained from command whoami) for normal run of Matlab or root if you want to run matlab with sudo (not recommended).
  • no installation window after sudo ./install → run xhost + first and than repeat
  • programs fail to run because permission of files in /home/username/.matlab/R202* → get to the folder .matlab (cd ~/.matlab) and run sudo chown -R username:username R202*, which will change group of the Matlab to your username (can be obtained from command whoami)
    • the * in R202* stands for your matlab version. If you downloaded it now, you will probably have R2022a

We will use Multisensorial calibration toolbox developed in our group. To install it, you can follow the instructions in README. The preferred (and the most easy) way to install is to download the installation script and run it in your Matlab (or Matlab online).

An introduction can be found in the presentation here. Everything else is described in README.

• Calibrate robot kinematic parameters from self-contact and self-observation information
• PDF with common mistakes is located here.