roc-swarm is a project that explores the coordination between drones/agents that were trained using NEAT (Neuro-Evolution of Augmenting Topologies) for flight control. The goal is to demonstrate how drones can autonomously collaborate to achieve a common objective in a simulated environment.
The clip displays two drones coordinating with each other to collect the rewards (which each drone should collect at least once)
I've tried it for two agents, and there is a lot to potentially develop like increasing multiple (both count-wise and trait-wise) agents, adding responses to different terrains, etc.
Clone and install dependencies
git clone https://github.com/sudo-boo/roc-swarm.git
cd roc-swarm
pip install -r requirements.txtRun the Simulation:
python roc_swarm_simulation.pyAdjust parameters such as drone behavior, environment setup, and simulation duration in config.ini.
This project is licensed under the MIT License - see the LICENSE file for details.