cfc_collision_ros

ROS package for CFC-based collision detection framework, with applications on motion planning.

Introduction

This is the ROS package for the collision detection algorithms using closed-form contact space (CFC) parameterization. It includes visualizations and applications on robot motion planning. Direct interface with OMPL is implemented.

• Paper: IEEE Robotics and Automation Letters (RA-L)
• Project page: https://chirikjianlab.github.io/cfc-collision/
• C++ core library: https://github.com/ChirikjianLab/cfc-collision/

Authors

Sipu Ruan, Xiaoli Wang and Gregory S. Chirikjian

Dependencies

• CFC-collision: Core C++ library for CFC-based collision detection, fetch as a submodule (see installation instructions).
• ROS Noetic (Ubuntu 20.04)
• Ceres solver (>= 2.0)

Required ROS packages

• MoveIt 1.0: Binary installation is recommended.
• panda_moveit_config
• ompl
• moveit_visual_tools
• eigen_conversions

Installation command

[sudo] apt install ros-noetic-moveit ros-noetic-panda-moveit-config ros-noetic-ompl ros-noetic-moveit-visual-tools ros-noetic-eigen-conversions

Installation

• Install required dependencies
• Generate catkin workspace and download the source code

mkdir -p ~/catkin_ws/src/
cd ~/catkin_ws/src/
git clone --recurse-submodules https://github.com/ChirikjianLab/cfc_collision_ros.git

• Build the package

cd ~/catkin_ws/
catkin build

Usage

Visualize collision using interactivity marker

Show contact points as described in MoveIt tutorial -- visualizing collisions.

roslaunch cfc_collision_ros moveit_visualize_collision.launch

Additional feature: when the object and robot are close to each other, a line segment connecting the closest witness points are shown.

Separated	In collision

Planning using OMPL

roslaunch cfc_collision_ros ompl_planning.launch

Note

• Please change the parameters for different robots (here) and environment types (here). Can also specify as

roslaunch cfc_collision_ros ompl_planning.launch robot_name:=panda_arm env_type:=sparse

• For real Panda robot, use ompl_planning_real.launch
• For end points in workspace (only supports panda_arm robot model), use ompl_planning_ee.launch. Before planning, please run IK solver launch file first (instructed below).

Reproduction of figures in paper

To reproduce figures in the paper (Figure 6), please use the following arguments when launching ompl_planning.launch:

Figure	Robot name	Environment type	Argument
Fig. 6(a)	Snake	Narrow	robot_name:=snake env_type:=narrow
Fig. 6(b)	Panda single arm	Dense	robot_name:=panda_arm env_type:=dense
Fig. 6(c)	Panda dual arm	Sparse	robot_name:=panda_dual_arm env_type:=sparse

Snake-like robot	Franka Emika Panda robot

Solve inverse kinematics

Given a set of SE(3) poses, solve inverse kinematics using the MoveIt feature (moveit::core::RobotState::setFromIK)

roslaunch cfc_collision_ros moveit_ik_solver.launch

Note

• For testing on real scenes, use moveit_ik_solver_real.launch

Status

Features

• Visualize collision and distance via interactive markers
• Direct interface with OMPL using CFC-based collision checker

Supported robots

• Panda arm (7 DOF)
• Panda dual-arm (14 DOF)
• Snake-like manipulator (15 DOF)

Available environment types

• Sparse
• Cluttered
• Narrow
• Dense
• Real lab experiments

Supported start/goal types

• Joint space (C-space): For all robots
• Workspace (end effector poses): Only for "panda_arm"

Reference

If you find our work useful in your research, please consider citing:

• S. Ruan, X. Wang and G. S. Chirikjian, "Collision Detection for Unions of Convex Bodies With Smooth Boundaries Using Closed-Form Contact Space Parameterization," in IEEE Robotics and Automation Letters, vol. 7, no. 4, pp. 9485-9492, Oct. 2022, doi: 10.1109/LRA.2022.3190629.

• BibTeX

@ARTICLE{ruan2022collision,
  author={Ruan, Sipu and Wang, Xiaoli and Chirikjian, Gregory S.},
  journal={IEEE Robotics and Automation Letters}, 
  title={Collision Detection for Unions of Convex Bodies With Smooth Boundaries Using Closed-Form Contact Space Parameterization}, 
  year={2022},
  volume={7},
  number={4},
  pages={9485-9492},
  doi={10.1109/LRA.2022.3190629}}