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 work has been accepted by RA-L 2022 and IROS 2022.

@ARTICLE{Sun_FISS_A_Trajectory_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}}

All Results above are obtained on an Intel i7-8700K CPU with 6 cores @3.70 GHz and 16 GB RAM

* are using Asynchronous collision checking module (with use_async = True in settings)

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
  • dynamic_reconfigure

Clone the repo and install dependencies:

# clone the repo
git clone https://github.com/SS47816/fiss_planner.git
cd fiss_planner

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

# build using `catkin_make` , or you can build with `catkin build`
catkin_make
# source 
source devel/setup.bash

Launch the Planner node by running:

# Launch nodes
roslaunch fiss_planner fiss_planner.launch

The latest instructions to run our code will be coming soon, since the LGSVL simulator is deprecated recently

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

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