mopy

Driver for Qualisys motion capture system using Python with the Qualisys Track Manager (QTM) API. Supports both local multiprocessed interface as well as ROS-based interface with corresponding launch files.

Brief

This repository provides a library for forwarding and filtering raw qualisys state measurements. The measurements provide position and attitude in form of a rotation matrix. (Unique and singularity-free representation of unit 2-sphere.) The rotation matrix is then converted to a quaternion and exposed to the desired interface.

The provided interfaces to Qualisys are either a multiprocessed environment for on-target implementation or ROS-based. Possible extensions using filtering algorithms are planned to be implemented. Skeletons are provided.

Usage

Essentially two distinct implementations are provided. First the local implementation without ROS using a multiprocessed environment and second a ROS-based implementation that publishes for each body name provided in the launch file argument model_names a separate topic in the private namespace of mopy. (Note: this can also be remapped to a desired topic -> a remap placeholder is provided.)

For example: Provided the launch command with model:=testname adding a remap to the launch file with <remap from="~testname" to="/global_namespace/desired_topic" />

• see also the documentation here

Raw Position and Attitude Measurement

• Execute ROS node
  • roslaunch mopy stream.launch
    • will create distinct topics for all models registered in current Qualisys configuration (server side)
  • roslaunch mopy stream.launch model:=modelname
    • will create distinct topic for modelnames (list of strings) without space in following syntax
      • roslaunch mopy stream.launch model:=modelname1,modelname2 filter:=True debug:=True
  • roslaunch mopy stream.launch model:=modelname filter:=True
    • will create distinct topic for modelnames and filter states (todo)
  • roslaunch mopy stream.launch model:=modelname filter:=True debug:=True
    • will create distinct topic for modelnames and filter states (todo) together with debug prints

Setup

• The following procedure is tested under Python 3.7
• Clone this repository in a catkin compliant workspace structure.
  • e.g. yourws/src/mopy
• Install dependencies
  • pip install -r requirements.txt
  • cd yourws/externals/transformations && python setup.py install
• build your workspace
  • cd yourws && catkin build
• source the workspace environment variables
  • e.g. source yourws/devel/setup.bash
• execute roslaunch and rosrun as detailed above.

Todo

• exit asyncio more gracefully and adjust
  • improve exit by restructuring and using signal.SIGINT events
• commit multiprocessed ROS-free implementation for QTM interaction
• include filter to estimate v, w

Contribution

Any form of contribution is very welcome and very much appreciated. For python scripts adhere to PEP8 and for ROS conventions REP standards (such as REP105 for frames).

For any questions feel free to raise an issue or for bug fixes to submit a pull request.

Contact: Philipp Rothenhäusler | E-mail: philipp.rothenhaeusler (gmail.com)

For Python:

• show PEP8 violations (and manual fix - recommended)

  pip install pycodestyle

  usage:

  pycodestyle --show-source --show-pep8 interfaces.py (show code)

  pycodestyle --statistics -qq interfaces.py (show common errors)

• automatically correct PEP8 violations

  pip install autopep8

  usage:

  autopep8 --in-place --aggressive --aggressive interface.py