superquadric-library

The goal of this project is to provide a Yarp-free library for computing and visualizing the superquadric representing an object and the relative grasping candidates for a generic robot.

This is a new design of the code included in the repositories:

• superquadric-model;
• superquadric-grasp;
• superquadric-grasp-demo.

Release 0.3

This release provides methods to:

• reconstruct the object model with single or multiple superquadric;
• compute grasping candidates for any robot;
• visualize superquadrics, planes, point clouds and grasping poses.

Some examples of outputs of the superquadric-lib are available here.

Relevant features:

• Only three dependencies: Ipopt,Eigen3 and VTK;
• Faster computation of grasping poses w.r.t the implementation of superquadric-grasp;
• High-level interface.

Dependencies

• Ipopt
• Eigen3
• VTK >= 8.1

How to build

Here are the instructions to build and install the library:

$ git clone https://github.com/robotology/superquadric-lib.git
$ cd superquadric-lib
$ mkdir build && cd build
$ cmake ..
$ make
$ [sudo] make install

By default, this command will build and install:

1. SuperquadricLibModel, that includes all the tools to play with superquadrics;
2. SuperquadricLibGrasp, to compute the grasping poses for an object represented with a superquadric (it requires 1.);
3. SuperquadricLibVis, to visualize everything (it requires 1. and 2.);
4. Superquadric-Pipeline-Single and Superquadric-Pipeline-Multiple, executables providing an example on how to compute single/multiple superquadric object models and the grasping pose starting from the object point cloud and show the results using the visualizer. They require 1 - 3.

Optional:

• Python bindings: Enabling the flag ENABLE_BINDINGS will generate the python bindings of SuperquadricLibModel, SuperquadricLibGrasp and SuperquadricLibVis. The following dependency is required to generate the python bindings:
• SWIG - 4.0.1, which can be installed with:

$ curl https://netcologne.dl.sourceforge.net/project/swig/swig/swig-4.0.1/swig-4.0.1.tar.gz | tar xvz
$ cd swig-4.0.1 && ./configure --prefix=<path-to-install-directory>
$ make -j4
$ make install

After the installation of the python bindings, you may need to add the superquadriclib/bindings to the PYTHONPATH in your .bashrc:

export PYTHONPATH=${PYTHONPATH}:${INSTALL_DIR}/lib/superquadriclib/bindings

• Yarp demo: Enabling the cmake flag YARP_EXE will compile also a Yarp-based-Demo, named Superquadric-Lib-Demo. This is a demo that uses the superquadric-lib to execute superquadric modeling and grasping on the iCub robot, iCub simulator and R1 robot.

The following dependencies are required for this demo:

• Yarp
• icub-main
• icub-contrib-common

An overview of the Superquadric-Lib-Demo is provided here.

How to link

Once the library is installed, you can link it using CMake by writing the following line of code in your project CMakeLists.txt:

find_package(SuperquadricLib 0.1.0.0 EXACT REQUIRED)
target_link_libraries(<target> SuperquadricLib::SuperquadricLib<tag>)

where <tag> can be:

• Model
• Grasp
• Vis

How to use the library

An example on how to use the library in your code is provided here.