pykin

pykin

Python Interface for the robot Kinematics library pykin

This library has been created simply by referring to ikpy.

Features

• Pure python library
• Support only URDF file
• Compute forward, inverse kinematics and jacobian, referred to the Introduction to Humanoid Robotics book.
• Check robot self-collision and collision between objects
• Plot robot kinematic chain and mesh

Installation

Requirements

You need a python-fcl package to do object collision checking.

• For Ubuntu, using apt

  sudo apt install liboctomap-dev

  sudo apt install libfcl-dev

• For Mac, First, Download the source and build it.

  • octomap

    git clone https://github.com/OctoMap/octomap.git

    $ cd octomap
    $ mkdir build
    $ cd build
    $ cmake ..
    $ sudo make
    $ sudo make install
    

  • fcl

    git clone https://github.com/flexible-collision-library/fcl.git

    Since python-fcl uses version 0.5.0 of fcl, checkout with tag 0.5.0

    $ cd fcl
    $ git checkout 0.5.0
    $ mkdir build
    $ cd build
    $ cmake ..
    $ sudo make
    $ sudo make install
    

Install Pykin

pip3 install pykin

Quick Start

You can see various examples in examples directory

• Robot Info

  You can see 7 robot info.

  baxter, sawyer, iiwa14, iiwa7, panda, ur5e, doosan

  $ cd examples
  $ python robot_info.py $(robot_name)
  # baxter
  $ python robot_info.py baxter
  # saywer
  $ python robot_info.py sawyer
  

• Forward kinematics

  You can compute the forward kinematics as well as visualize the visual or collision geometry.

  $ cd examples/forward_kinematics
  $ python robot_fk_baxter_test.py
  

  visual	collision

• Inverse Kinematics

  You can compute the inverse kinematics using levenberg marquardt(LM) or newton raphson(NR) method

  $ cd examples/inverse_kinematics
  $ python robot_ik_baxter_test.py
  

• Sampling based Inverse Kinematics

  You can compute the inverse kinematics using geometric-aware bayesian optimization(GaBO) method

  For more detailed information, check GaBO module

  $ cd examples/inverse_kinematics
  $ python robot_ik_gabo_test.py
  

• Collision check

  The below images show the collision result as well as visualize robot using trimesh.Scene class

  $ cd examples/trimesh_renders
  $ python sawyer_render.py
  

  trimesh.Scene	Result

Visualization

You can see visualization using matplotlib library or trimesh.Scene class.

• Visualize simple urdf using matplotlib

  ur5e	sawyer	iiwa14	panda

• Visualize visual geometry using matplotlib

  ur5e	sawyer	iiwa14	panda

• Visualize collision geometry using matplotlib

  ur5e	sawyer	iiwa14	panda

• Visualize mesh about visual/collision geometry using trimesh.Scene class