Homotopy-aware Multi-agent Navigation via Distributed Model Predictive Control

This repository provides an algorithm implementation of the paper Homotopy-aware Multi-agent Navigation via Distributed Model Predictive Control. The details of this work are as follows:

• Authors: Haoze Dong, Meng Guo, Chengyi He and Zhongkui Li
• Paper: to be released
• Video: YouTube

Dependencies Installation

The system is tested on Ubuntu 20.04 with Python 3.

Install Required Python Packages

pip install numpy scipy pybind11==2.11.1 rospkg==1.5.0 setuptools==70.0.0

Install Eigen

sudo apt-get install libeigen3-dev
sudo ln -s /usr/include/eigen3/Eigen /usr/include/Eigen

Install ROS

Follow the official ROS installation guide: ROS Installation

Install Acados

Refer to the Acados documentation to install Acados and its Python interface: Acados Python interface

Note: t_renderer binaries may need to be built manually from tera_renderer

Install pcl for obstacle visualization

sudo apt install python3-pcl
sudo apt install ros-noetic-pcl-ros

Install tmux

sudo apt-get install tmux

Build Instruction

Build up the workspace

git clone https://github.com/HauserDong/HomoMPC
cd HomoMPC
catkin_make
source devel/setup.bash

Build GJK & Homotopy-aware Path Planning algorithm code

Note: Before running the command below, ensure you are in the root of your workspace.

source easy_build.sh

Running the simulation

Important: Always source the workspace devel/setup.bash script when opening a new terminal. Hint: you can add it to your .bashrc for convenience.

Launch the System

roslaunch bring_up bring_up.launch 

Run the planner

python3 src/local_planner/planner/scripts/launch.py

To terminate the tmux windows, try tmux kill-server in your command line.

Publish targets

python3 src/local_planner/planner/scripts/publish.py

Acknowledgements

The implementation of this project is founded upon the following packages.

IMPC-OB: https://github.com/PKU-MACDLab/IMPC-OB

ASAP: https://github.com/CYDXYYJ/ASAP

acados: https://github.com/acados/acados

HPSP: https://github.com/HuangJingGitHub/HPSP

tuw_multi_robot: https://github.com/tuw-robotics/tuw_multi_robot

openGJK: https://www.mattiamontanari.com/opengjk/