Run article tendon gripper generate

• install env (see Installation in development mode)
• activate rostok env by enter conda activate rostok
• run pipline python app\app_new_mcts_parallel.py

Structure tendon gripper pipiline

• rostok\library\rule_sets\ruleset_simple_fingers.py -- contains a set of rules
• rostok\library\obj_grasp\objects.py -- contains a set of objects
• app\hyperparameters.py -- contains a set of hyperparameters

How to change grasp objects

Modify object grasp_object_blueprint in python app\app_new_mcts_parallel.py. Use predefined function from rostok\library\obj_grasp\objects.py.

Rostok

Rostok is an open source Python framework for generative design of linkage mechanisms for robotic purposes. It provides a framework to describe mechanisms as a graph, set an environment, perform simulation of generated mechanisms, get a reward as a quantitative value of the generated design, and search for the best possible design.

A user can utilize the entire framework as a pipeline to generate a set of suboptimal designs, or utilize the modules and submodules as independent parts. The framework allows to implement custom generative rules, modify search and optimization algorithms.

Currently the framework allows to perform co-design of open chain linkage mechanisms. Co-design consists in simultaneously searching for the mechanical structure and the trajectories of the robot to get the best possible performance.

Prerequisites

• Anaconda3
• Usage of the Docker reqires installation of Х-server for Windows https://sourceforge.net/projects/vcxsrv/

Installation in development mode

To modify the modules of the Rostok framework a user should install it in development mode:

• Create the environment using conda env create -f environment.yml
• Activate the environment rostok
• Install the package in development mode pip3 install -e .

Documentation

The description of the project and tutorials are available at project website.

Acknowledgments

Affiliation

The framework was developed in ITMO University.

Supported by

The study is supported by the Research Center Strong Artificial Intelligence in Industry of ITMO University as part of the plan of the center's program: Development and testing of an experimental prototype of a library of strong AI algorithms in terms of generative and interactive design of planar mechanisms of anthropomorphic gripping devices and robotic hands

Developers

• Ivan Borisov - researcher
• Kirill Zharkov - team leader
• Yefim Osipov - developer
• Dmitriy Ivolga - developer
• Kirill Nasonov - developer
• Mikhail Chaikovskii - developer
• Sergey Kolyubin - chief scientist

Contacts

• Ivan Borisov [email protected] for scientific aspects of the project
• Kirill Zharkov [email protected] for technical questions of the project
• Sergey Kolyubin [email protected] for collaboration aspects

Citation

GOST:

• I. I. Borisov, E. E. Khomutov, D. V. Ivolga, N. A. Molchanov, I. A. Maksimov and S. A. Kolyubin, "Reconfigurable Underactuated Adaptive Gripper Designed by Morphological Computation," 2022 International Conference on Robotics and Automation (ICRA), 2022, pp. 1130-1136, doi: 10.1109/ICRA46639.2022.9811738.

Bibtex:

• @inproceedings{borisov2022reconfigurable, title={Reconfigurable underactuated adaptive gripper designed by morphological computation}, author={Borisov, Ivan I and Khomutov, Evgenii E and Ivolga, Dmitriy V and Molchanov, Nikita A and Maksimov, Ivan A and Kolyubin, Sergey A}, booktitle={2022 International Conference on Robotics and Automation (ICRA)}, pages={1130--1136}, year={2022}, organization={IEEE} }