Skip to content

course project for RO47005 Planning and Decision Making

License

Notifications You must be signed in to change notification settings

smoggy-P/kPRM-star-planner

 
 

Repository files navigation

Contributors Forks Stargazers Issues MIT License

An implementation of k-PRM path planner with corridor-based trajectory optimization

This is course project of Group 17 for RO47005 Planning and Decision Making in TU Delft. We implemented k-PRM for front-end path searching and an iterative method for back-end corridor-based trajectory optimization.

01

1. Our contribution

  • We implement occupancy map generation from scratch (please check src/planner/map_server/).
  • We implement PRM planner and k-d tree from scratch .
  • We implement A-star method for graph search from scratch (please check src/planner/prm_planner/).
  • We implement classic minimum-snap trajectory optimization (both optimization solution and closed-form solution) from scratch (please check src/planner/traj_opt/).
  • We implement corridor-based trajectory optimization from scratch (please check src/planner/traj_opt/). We use OSQP_Interface to solve QP problem. We also use Root-Finder to find tangent point of polynomial trajectory.
  • The simulation environment (including simulator and controller) we used is cloned from Fast-Planner
  • We use DecompROS for safe flight corridor generation directly from occupancy map.
  • We use mockamap with tiny modification to randomly generate point clouds and publish only once.
  • We compared our PRM planner with RRTs

2. Build a demo

Step 1. Install Armadillo, which is required by uav_simulator package.

sudo apt-get install libarmadillo-dev

Step 2. Install OSQP, which is required by traj_opt package.

Step 3. clone this repo

cd ~/<your_catkin_ws>/src
git clone https://github.com/edmundwsy/RO47005-PDM-Final.git
cd ..

Step 4. Build the project

catkin_make

3. Run a simple demo

Load configuration files

source devel/setup.bash

3.1 Demo in random forest environment

Run following command to launch our project in rviz

roslaunch planner simulation.launch

Use 2D Nav Goal in rviz to publish a goal point

3.2 Demo in 3D Maze environment

Run following command to launch our project in rviz

roslaunch planner maze3d.launch

3.3 Compare with RRT

Run following command to launch rrt_planner

roslaunch rrt_planner simulation_with_map.launch

Structure

├── ./planner
│   ├── ./planner/prm_planner   <our implementation>
│   ├── ./planner/map_server    <our implementation>
│   ├── ./planner/rrt_planner
│   ├── ./planner/traj_opt      <our implementation>
│   └── ./planner/Utils
└── ./uav_simulator
    ├── ./uav_simulator/fake_drone
    ├── ./uav_simulator/local_sensing
    ├── ./uav_simulator/map_generator
    ├── ./uav_simulator/mockamap
    ├── ./uav_simulator/so3_control
    ├── ./uav_simulator/so3_quadrotor_simulator
    └── ./uav_simulator/Utils

License

Distributed under the MIT License. See LICENSE for more information.

Contact

Siyuan Wu - S.Wu-14@student.tudelft.nl Moji Shi - M.Shi-5@student.tudelft.nl Ranbao Deng - R.Deng@student.tudelft.nl Liangchen Sui - L.Sui@student.tudelft.nl

About

course project for RO47005 Planning and Decision Making

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 80.8%
  • CMake 7.3%
  • Python 6.5%
  • C 2.2%
  • Cuda 2.0%
  • CSS 0.6%
  • Other 0.6%