Emergent Coordination through Game-Induced Nonlinear Opinion Dynamics

Table of Contents

1. About The Project
2. Dependencies
3. Example
4. License
5. Contact
6. Paper

About The Project

This repository implements a receding-horizon planning framework for multi-agent emergent coordination tasks based on Game-Induced Nonlinear Opinion Dynamics (GiNOD), and contains a simulated toll station coordination example involving two autonomous vehicles. The repo is primarily developed and maintained by Haimin Hu, a PhD student in the Safe Robotics Lab.

The ILQ Game implementation is inspired by David Fridovich-Keil's ilqgames and leverages JAX for real-time computation performance based on automatic differentiation and just-in-time (JIT) compilation.

Dependencies

This repo depends on the following packages:

1. jax=0.4.19
2. jaxlib=0.4.16
3. numpy=1.25.2
4. python=3.9.18
5. yaml=0.2.5
6. casadi=3.6.4

Example

The following scripts reproduce the results in our paper.

1. L0-QMDP with homogeneous cost weights

    python3 test_L0_two_car_trial_1.py

2. L0-QMDP with heterogeneou cost weights

    python3 test_L0_two_car_trial_2.py

3. L1-QMDP with heterogeneou cost weights

    python3 test_L1L0_two_car.py

License

Distributed under the BSD 3-Clause License. See LICENSE for more information.

Contact

Haimin Hu - @HaiminHu - haiminh@princeton.edu

Paper

If you find this repo helpful, please consider citing our paper.

@inproceedings{hu2023emergent,
      title={Emergent Coordination through Game-Induced Nonlinear Opinion Dynamics}, 
      author={Haimin Hu and Kensuke Nakamura and Kai-Chieh Hsu and Naomi Ehrich Leonard and Jaime Fernández Fisac},
      booktitle={2023 62nd IEEE Conference on Decision and Control (CDC), to appear},
      year={2023},
}