Gym Flock

Dependencies

• OpenAI Gym 0.11.0
• Python 3 (Python 2 doesn't work)
• AirSim (optional)

To install

1. Clone this repository
2. cd gym-flock
3. pip3 install -e .

Quick test

In the gym-flock folder:

1. python3 test.py

To use

Include the following code in your Python script:

import gym  
import gym_flock 
env = gym.make("Coverage-v0")` 

and then use the env.reset() and env.step() for interfacing with the environment as you would with other OpenAI Gym environments. These implementations also include a env.controller() function that gives the best current set of actions to be used for imitation learning.

The learning code for the flocking task can be found here

The learning code for the spatial coverage tasks can be found here

Using AirSim

Install AirSim according to instructions here

Move the settings.json file from gym-flock/gym_flock/envs/airsim/ to ~/Documents/AirSim/ Try changing the configuration to change the number of robots. The default number of robots is 5.

Launch AirSim

Launch test code with environment name FlockingAirsimAccel-v0. The code for this environment can be found here The environment will read the configuration file to identify the number of robots in the simulation.

Citing the Project

To cite the flocking environment in publications:

@inproceedings{tolstaya2020learning,
  title={Learning decentralized controllers for robot swarms with graph neural networks},
  author={Tolstaya, Ekaterina and Gama, Fernando and Paulos, James and Pappas, George and Kumar, Vijay and Ribeiro, Alejandro},
  booktitle={Conference on Robot Learning},
  pages={671--682},
  year={2020}
}

To cite the spatial coverage code in publications:

@misc{tolstaya2020multirobot,
      title={Multi-Robot Coverage and Exploration using Spatial Graph Neural Networks}, 
      author={Ekaterina Tolstaya and James Paulos and Vijay Kumar and Alejandro Ribeiro},
      year={2020},
      eprint={2011.01119},
      archivePrefix={arXiv},
      primaryClass={cs.RO}
}