Week 20. Networking and learning in the Gym-Pyxies tool.

2 minute read

To Do

  • Z.
  • Learning about Gym-Pyxies.


This beginning of the year has brought progress in the progress of the work. Small goals have been achieved that create a foundation with which to begin building the core work, teaching a vehicle to follow a line using Reinforcement Learning.

1. Train a network with the Vanessa’s dataset

The neural network of Vanessa Fernandez’s work has been trained on a dedicated machine resulting in a functional model that has been able to run successfully.

As a point of review for next week is to check the TensorFlow facilities to verify that the GPU is being used during training as it has taken too long in previous training.

2. Learning about Gym-Pyxies

Finally all the Player/Stage software has been installed.

Many things have been learned in this process. During the installation of both libraries there were problems with compilation:

  • Player: Compilation error on Ubuntu 18.04. The following library is missing:

    /usr/bin/ld: cannot find -lgeos

    To solve this, install the following library:

    sudo apt install libgeos++-dev

    More information about the error in this link.

  • Stage: Compilation error on Ubuntu 18.04. The following library is missing:

    Could NOT find FLTK (missing: FLTK_LIBRARIES)

    To solve this, install the following depencences:

    sudo apt install fltk1.3-dev
    sudo apt install libxft-dev
    sudo apt install libxinerama-dev

    More information about the error in this link.

Error in execution (Player). The terminal returns the following error:

Player v.3.1.1-dev

* Part of the Player/Stage/Gazebo Project [http://playerstage.sourceforge.net].
* Copyright (C) 2000 - 2013 Brian Gerkey, Richard Vaughan, Andrew Howard,
* Nate Koenig, and contributors. Released under the GNU General Public License.
* Player comes with ABSOLUTELY NO WARRANTY.  This is free software, and you
* are welcome to redistribute it under certain conditions; see COPYING
* for details.

error   : Failed to load plugin stageplugin.
error   : libtool reports error: file not found
error   : plugin search path: /usr/local/share/stage/worlds:/usr/local/share/stage/worlds:/usr/local/lib/player-3.1/:/usr/local/lib/
error   : failed to load plugin: stageplugin
error   : failed to parse config file /usr/local/share/stage/worlds/simple.cfg driver blocks

To solve this, change the environment variable to this one:

export LD_LIBRARY_PATH=/usr/local/lib/:$LD_LIBRARY_PATH

The library is in /usr/local/lib/ instead of /usr/local/lib64/. For more information, visit this link.

Once all the issues were solved, the example looks like this.

Example of Player/Stage

With this we already have all the tools prepared to be able to go deeper into this software and create our first Deep Reinforcement Learning project. The task for next week is to go deeper into the library.


Go deeper into the Player/Stage library by running more examples and understanding the relationship between them and communication.


Patience and perseverance. Reading, researching, searching, asking…in the end you get small steps that together give you small objectives that together get you to the final point. This will have to be the mantra of every day until the end of the work as well as applying it to the personal and professional life.