NU-IDEAS-Lab / patrolling_sim

Fork and reimplementation of https://github.com/davidbsp/patrolling_sim

patrolling_sim

This began as a fork of David Portugal's patrolling_sim, but has evolved into a rewrite of the original. This version makes use of the Grex Machina multi-agent framework and ROS 2, allowing for far more capabilities than the original patrolling_sim.

Papers

This repository is used in multiple papers. See below for the revisions associated with specific submissions:

• A. Goeckner, Y. Sui, N. Martinet, X. Li, and Q. Zhu, “Graph Neural Network-based Multi-agent Reinforcement Learning for Resilient Distributed Coordination of Multi-Robot Systems,” under review.

  • Commit 357e9a47064d67a6b6f66f41f73424933f86e345

• A. Goeckner, X. Li, E. Wei, and Q. Zhu, “Attrition-Aware Adaptation for Multi-Agent Patrolling.” arXiv, Jan. 12, 2024. doi: 10.48550/arXiv.2304.01386.

  • Commit 36e474b734c9682798e0f0260ba9fda632c3b6ca

Prerequisites

First install all prerequisites of the Grex Machina repository.

Then, perform the following:

• Install Python dependencies.

  pip install networkx[default] numpy matplotlib zarr

Installation

Once prerequisites are installed, perform the following:

1. Use Git to download the patrolling sim repository.

   git clone --recurse https://github.com/NU-IDEAS-Lab/patrolling_sim.git

2. Change to the patrolling_sim folder:

   cd ./patrolling_sim/

3. Source the correct ROS 2 installation.

   source /opt/ros/humble/setup.bash

4. Build the patrolling sim using Colcon, the ROS 2 build system.

   colcon build

5. Follow instructions below to run the simulation.

Operation Instructions

To operate the simulator, perform the following steps:

1. Change to the patrolling_sim root directory:

   cd ./patrolling_sim/
   

2. Source the installed packages, plus the Gazebo simulator, and set environment variables.

   source ./install/setup.bash
   source /usr/share/gazebo/setup.sh
   export RMW_IMPLEMENTATION=rmw_cyclonedds_cpp
   

3. Run the simulation:

   ros2 launch patrolling_sim simulation.launch.py agent_count:=2
   

   • Note: To view available launch arguments, add a -s flag to the end of the command.