Kineverse

Kineverse is a framework for building models of articulated objects/robots using symbolic mathematical expressions and constraints. It is meant to be used in a ROS environment and features a client-server implementation which allows articulation models to be modified and updated across multiple nodes in a running ROS system.

Version 2

Kineverse has gotten (or is in the progress of getting) a full re-write as kv-lite, as this version made a good first step, but is very difficult to use. The second version is meant to be more streamlined and tries as much as possible to behave like numpy to be easier to approach. This repository will remain for the sake of posterity.

Installation

Kineverse depends on a number of different libraries, some of which need to be installed manually. The installation has been tested for Ubuntu 20.04 and ROS Noetic.

Kineverse

Go to a ROS workspace and clone Kineverse.

cd ros_ws/src  # Go to a ROS workspace

git clone https://github.com/ARoefer/kineverse.git
git clone https://github.com/ARoefer/kineverse_msgs.git

Dependencies

Kineverse depends on a number of external packages, some of which can be installed using pip. From within the kineverse directory, run:

pip install -r requirements.txt

QP-Oases

For the time being, Kineverse contains a small, QP-based motion generation module. This module requires the QP-Oases ROS package to be installed.

cd ros_ws/src  # Go to a ROS workspace
git clone https://github.com/SemRoCo/qpOASES.git
cd qpOASES
git checkout noetic

Once you are done cloning the package, you can rebuild and re-source your workspace.

Installing Bullet

Kineverse includes a small geometry query system which uses a custom wrapper for the Bullet Physics Engine. Unfortunately, the installation has to be done from source. The kineverse package includes a build_pybullet.sh script which can be run to perform the entire installation automatically. The script expects a directory as argument to which it will clone both pybind11 and bullet3. It will build and install both from there. The script will ask for sudo rights as a part of the process.

# Ensure you have the Python 3 development tools
sudo apt install python3-dev 

# Run installation
./build_pybullet.sh /my/awesome/library/dir
# Wait...

If everything worked, you should be able to import the wrapper in a Python environment.

import betterpybullet as bpb

If that does not work, try restarting your system.