rapid-transport

rapid-transport is the accompanying source code of the paper "Critically fast pick-and-place with suction cups" submitted to ICRA 2019. rapid-transport implements a motion planning pipeline that produces fast and robust motion for object transportation.

A video of the experiments of this pipeline can be found below.

You are welcome to clone this package and experiment with the code. If you have any questions pertaining to the code, do raise an issue via the Github Issue Tracking system. See below for some instructions on installing and running.

Installation

rapid-transport has two main dependencies: toppra and OpenRAVE.

Here are instructions on installing both packages for your reference:

# install openrave
git clone https://github.com/crigroup/openrave-installation
cd openrave-installation
./install-dependencies.sh
./install-osg.sh
./install-fcl.sh
./install-openrave.sh

# install toppra
git clone https://github.com/hungpham2511/toppra && cd toppra/
git checkout 86cd270  # preferred version of toppra, you can try the lastest one but not guaranteed to work.
pip install -r requirements.txt --user
python setup.py install --user

Now you can run rapid-transport

git clone https://github.com/hungpham2511/rapid-transport && cd rapid-transport
pip install -r requirements.txt --user
python setup.py develop --user

For more information on toppra, see this page.

Experiment with the code

rapid-transport provides two executables that you might be interested in: i) generate_contact_constraint.py and ii) transport.paper pick-demo.

Running generate_contact_constraint.py

cd <rapid-transport>dir
python scripts/generate_contact_constraint.py

This executable performs the contact stability constraint approximation procedure described in the paper, and outputs the constraint as a numpy archive.

If you are just trying out, check data/contacts.yaml and data/contact_data/analytical_rigid1234.npyz. This is a pre-generated constraint that is used in the paper.

Running pick-demo

transport.paper pick-demo --scene "scenarios/exp2.scenario.yaml"

This executable runs the demo discussed in the paper. You will need OpenRAVE to run this one.

Experiment with the real robot

roslaunch denso_control rc8_ros_driver.launch rate:=125
# or to run gazebo simulation
roslaunch denso_gazebo denso_vs060.launch

roslaunch denso_control joint_trajectory_controller.launch namespace:=denso

transport.paper pick-demo --scene "scenarios/dimdatap2.scenario.yaml" -d 0.9 -e 3 -m RUN