A novel Hybrid Decoupled Visual Servoing (HDVS) method implemented in C++ and ROS for Franka Panda robots.
git clone https://github.com/frankaemika/franka_ros.git
git clone https://github.com/erdalpekel/panda_simulation.git
git clone https://github.com/neka-nat/gazebo_domain_randomization.git
Unzip the visp_ros.zip file inside this repository to your src folder in your ros workspace. The zip file contains visp_ros and vision_visp ros nodes based on the VISP library GitHub Pages.
To run the simulation without domain randomization:
roslaunch hdvs_launch simulation-wo-DM.launch
To run the simulation with domain randomization:
roslaunch hdvs_launch simulation.launch
Inside the src folder of hdvs_code, you'll find codes dedicated to various types of visual servoing. Notably, the Pure_vs code is designed for universal use across different robot types. This code calculates the generated velocity from each visual servoing method and publishes the resulting data on the assigned ROS topic.
The C++ folder contains the code intended for execution on the real robot, utilizing the libfranka and VISP libraries.
The findings of this repository were published in:
@article{rastegarpanah2022optimized,
title={Optimized hybrid decoupled visual servoing with supervised learning},
author={Rastegarpanah, Alireza and Aflakian, Ali and Stolkin, Rustam},
journal={Proceedings of the Institution of Mechanical Engineers, Part I: Journal of Systems and Control Engineering},
volume={236},
number={2},
pages={338--354},
year={2022},
publisher={SAGE Publications Sage UK: London, England}
}
@article{rastegarpanah2021improving,
title={Improving the manipulability of a redundant arm using decoupled hybrid visual servoing},
author={Rastegarpanah, Alireza and Aflakian, Ali and Stolkin, Rustam},
journal={Applied Sciences},
volume={11},
number={23},
pages={11566},
year={2021},
publisher={MDPI}
}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent