Skip to content

[ICRA2024] Stein Variational Guided Model Predictive Path Integral Control: Proposal and Experiments with Fast Maneuvering Vehicles

License

Notifications You must be signed in to change notification settings

kohonda/proj-svg_mppi

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

25 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Stein Variational Guided Model Predictive Path Integral Control (SVG-MPPI)

This package includes ROS implementation of Stein Variational Guided Model Predictive Path Integral Control: Proposal and Experiments with Fast Maneuvering Vehicles presented in ICRA 2024.

agile_driving.mp4

Overview

Tested Environment

  • Ubuntu 20.04, 22.04, and 24.04
  • Docker
Docker Installation

Installation guide

# Install from get.docker.com
curl -fsSL https://get.docker.com -o get-docker.sh
sudo sh get-docker.sh
sudo groupadd docker
sudo usermod -aG docker $USER

Quick Start with Docker

1. Build Docker Image

make docker_build

2. Launch Simulator and Controllers

Launch simulator in a new terminal

cd proj-svg_mppi
./script/launch_simulator.sh

Launch controllers in another terminal

cd proj-svg_mppi
# Enter Docker container with GUI support
make bash
# Inside container, launch controllers
./script/launch_controllers.sh

You can change the all MPPI parameters and settings in the yaml file

simulation.mp4

Run on Native System

Requirements

  • Ubuntu 20.04
  • ROS Noetic

1. Install Dependencies

cd proj-svg_mppi
make setup

2. Build the Project

cd proj-svg_mppi
make build

3. Launch Simulator and Controllers

Launch simulator in the Docker container

cd proj-svg_mppi
./script/launch_simulator.sh

Launch controllers in another terminal

cd proj-svg_mppi
./script/launch_controllers.sh 

Evaluation

You can reproduce all simulation results in the paper by one command:

cd proj-svg_mppi/script
./eval_all.sh

Or, You can evaluate a fixed parameters by this command:

cd proj-svg_mppi/script
./eval.sh

Note: The evaluation is used asynchronous simulation using ROS. So, the results can be slightly changed even if all seeds are fixed.

Citation

@inproceedings{honda2024stein,
  title={Stein Variational Guided Model Predictive Path Integral Control: Proposal and Experiments with Fast Maneuvering Vehicles},
  author={Honda, Kohei and Akai, Naoki and Suzuki, Kosuke and Aoki, Mizuho and Hosogaya, Hirotaka and Okuda, Hiroyuki and Suzuki, Tatsuya},
  booktitle={2024 IEEE International Conference on Robotics and Automation (ICRA)},
  pages={7020--7026},
  year={2024},
  organization={IEEE}
}

About

[ICRA2024] Stein Variational Guided Model Predictive Path Integral Control: Proposal and Experiments with Fast Maneuvering Vehicles

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published