METURONE-Swarm is a decentralized controller for homogeneous multi-agent systems. With METURONE-Swarm, you can control up to 12(it is tested with 12 agents, if you test it with more agents, please let us know!) agents in an obstacle-free environment. Current features are formation constructing, trajectory following, rotating the swarm around its center and dividing the group into any number of groups, each with a different mission.
First, clone the repository.
git clone https://github.com/METUrone/METURONE-Swarm.gitThen install the required packages with
pip3 install requirements.txtThat's it! However, if you will use ROS to communicate with your robots, you also need to install ROS from https://wiki.ros.org/ROS/Installation.
Although the system allows using with any kind of agent, current code only supports Crazyflie agents with OptiTrack systems. To adapt it to your system, you should rewrite poses.py, get the position and orientation information somehow and print it. We are aware that this is not performance-friendly, and will change it soon. There are also Crazyflie-specific code in commander.py, you may need to change or delete it according to your needs.
You may also implement a different algorithm very easily with METURONE-Swarm. The only thing you should do is implementing calculate_speed and CalculateCollisionSpeed functions in uav.py.
After you adapt the system to your agents, you only run below command.
python3 gui3.py
In the interface you can watch the agents from the left panel and right below panel, and give commands from the right top panel on-the-fly or before the mission.
Feel free to open issues and pull requests, and you can also read our report for this project METURONE-Swarm.pdf. However, it is only in Turkish for now.