ClimbLab - Climbing robot Matlab simulator

Author(s) and maintainer(s): Space Robotics Lab. Climbing Robotics Team

The climbing simulator is being developed by the Climbing Robotics Team in Space Robotics Laboratory at Tohoku University, Japan.

The source code is released under a BSD 3-Clause license.

Overview

This simulator wraps up functions for:

• [x] Dynamic analysis and simulation of articulated multi-body systems with a floating base (This is performed by SpaceDyn [1])
• [x] Design of legged robotic system
• [x] Environment demonstration (different terrain shape, inclination, and gravity)
• [x] Visualization (Robot, map, support polygons, stability criterion, and time-history of any state variables)

Simulation examples

1) Standard gait on uneven terrain	2) Optimal foothold selection on descrete grippable points [1]

3) Stability region visualization [2]	4) Topographically salient region detection and non-periodic foothold selection [3]

Publication

If you use this simulator in an academic context, please cite the following publication.

K. Uno, W. F. R. Ribeiro, Y. Koizumi, K. Haji, K. Kurihara, W. Jones, and K. Yoshida "ClimbLab: MATLAB Simulation Platform for Legged Climbing Robotics", Proceedings of the 24th International Conference on Climbing and Walking Robots and the Support Technologies for Mobile Machines (CLAWAR), 2021.

@inproceedings{uno2021climblab,
  title={ClimbLab: MATLAB Simulation Platform for Legged Climbing Robotics},
  author={Kentaro Uno and Warley F. R. Ribeiro and Yusuke Koizumi and Keigo Haji and Koki Kurihara and William Jones and Kazuya Yoshida},
  booktitle={Proceedings of the 24th International Conference on Climbing and Walking Robots and the Support Technologies for Mobile Machines (CLAWAR)},
  pages={TBD},
  year={2021}
}

Requirements

We confirmed the code is working with:

• Matlab 2018b or higher
• SpaceDyn - a MATLAB Toolbox for Space and Mobile Robots (version 2, release 0)
  • SpaceDyn toolbox v2r0 is also installed by cloning this repository (under lib/spacedyn_v2r0)

Build and Run

• Download the files from this repository
• Open MATLAB
• Select ClimbLab/ as the current folder for MATLAB
  • You should see config/, dat/, docs/, lib/ and src/ directories under it
• Setup the parameters as you want by changing the variable config.
  • Note: If you select USER configuration type, follow instructions described in the config/USER/config_USER_param_template.m.
• Run src/climb_main.m file (Note: MATLAB might ask you to change folder or add path. Choose to "add path")

References

Our recent works using ClimbLab simulator are as follows:

[1] Kentaro Uno et al., "Gait Planning for a Free-Climbing Robot Based on Tumble Stability", Proceedings of the 2019 IEEE/SICE International Symposium on System Integration (SII), Paris, France, 2019, pp. 289-294, doi: 10.1109/SII.2019.8700455.

[2] Warley F. R. Ribeiro et al., "Dynamic Equilibrium of Climbing Robots based on Stability Polyhedron for Gravito-Inertial Acceleration", Proceedings of the 23rd International Conference on Climbing and Walking Robots and the Support Technologies for Mobile Machines, Moscow, Russian Federation, 2020, pp. 297-304, doi: 10.13180/clawar.2020.24-26.08.18.

[3] Kentaro Uno et al., "Non-Periodic Gait Planning Based on Salient Region Detection for a Planetary Cave Exploration Robot", Proceedings of the International Symposium on Artificial Intelligence, Robotics and Automation in Space, remote conference, 2020, No. 5027.

FAQ

Please contact us by sending an email to the following address if there is any question or suggestion.

srl-limb[at]grp.tohoku.ac.jp

Release note

• Version 1.0: released on 18th Sept. 2020.
• Version 2.0: released on 26th Oct. 2020. This version includes:
  • topographically salient region detection algorithm [3]
  • non-periodic gait planner [3]
  • gripper detachment evaluation
• Version 3.0: released on 27th Aug.2021. This version includes
  • Camera sensing
  • New robot models
  • More analysis scripts