Skip to content

[RA-L 2022] FISS: A Trajectory Planning Framework using Fast Iterative Search and Sampling Strategy for Autonomous Driving

License

Notifications You must be signed in to change notification settings

SS47816/fiss_planner

Repository files navigation

FISS Planner

[RA-L 2022] FISS: A Trajectory Planning Framework using Fast Iterative Search and Sampling Strategy for Autonomous Driving

Ubuntu ROS C++ CodeFactor GitHub Repo stars GitHub Repo forks

This is an improved framework based on the framework used in frenet_optimal_planner. We utilize historical planning results as prior information in heuristics and introduce an iterative search-generate-evaluate strategy to look for the optimal trajectory candidate. Compared to the current frameworks, our method can significantly reduce the number of computationally expensive operations spent during the planning and achieve 2-6 times faster.

Our paper has been accepted by RA-L 2022 and IROS 2022.

@article{Sun_FISS_2022,
  author={Sun, Shuo and Liu, Zhiyang and Yin, Huan and Ang, Marcelo H.},
  journal={IEEE Robotics and Automation Letters},
  title={FISS: A Trajectory Planning Framework Using Fast Iterative Search and Sampling Strategy for Autonomous Driving},
  year={2022},
  volume={7},
  number={4},
  pages={9985-9992},
  doi={10.1109/LRA.2022.3191940}
}

Video

video

cover_image

Updates

  • [14 Jul 2023] An improved new version of the method, FISS+, is now available at https://github.com/SS47816/fiss_plus_planner
  • [19 Oct 2022] Major updates in the documentation. Now you should be able use this repo out-of-the box.

Dependencies

Our package is only based on standard ROS pkgs, with no other external dependencies:

  • C++11 above
  • CMake: 3.0.2 above
  • Eigen (included)
  • ROS Packages:
    • roscpp
    • rospy
    • tf
    • tf2_ros
    • std_msgs
    • nav_msgs
    • geometry_msgs
    • autoware_msgs
    • visualization_msgs
    • tf2_geometry_msgs
    • dynamic_reconfigure

Installation

To use this package, you will need to create a catkin_ws first. Details please see the ROS official tutorial.

# locate your catkin workspace (assuming ~/catkin_ws here)
cd ~/catkin_ws/src

# clone the lgsvl_utils repo so that you can use this planner with the lgsvl simulator
git clone https://github.com/SS47816/lgsvl_utils.git
# clone this repo
git clone https://github.com/SS47816/fiss_planner.git

cd ..

# install dependencies
rosdep install --from-paths src --ignore-src -r -y

# build
catkin_make
# source
source devel/setup.bash

Usage

  1. Install the LGSVL simulator by following this guide
  2. Set up your LGSVL simulator and launch the lgsvl_utils nodes by following the guide
  3. Launch the FISS planner nodes by running:
# Launch nodes
roslaunch fiss_planner fiss_planner.launch

The local planner is now waiting for the global route to be published so that it can start planning. 4. You may adjust any config parameters you like in the dynamic_reconfigure window. 5. In the RVIZ window, use the 2D Nav Goal tool to selet a reachable goal point on the road as the global goal. The global planner will immediately plan a global route for you. And you will soon see the local planner starts planning. 6. Press the green A button on your joystick to enter the autonomous mode. 7. Now you should be able to see your ego vehicle moving. Have fun!

Contribution

You are welcome contributing to the package by opening a pull-request

We are following: Google C++ Style Guide, C++ Core Guidelines, and ROS C++ Style Guide

License

Our fiss_planner ROS package is licensed under Apache License 2.0

The included Eigen Library follows its own Mozilla Public License v. 2.0