Ricardo A. Téllez

This section contains all the material that I have generated during my research related to Aibo programming. You can use it for free but you must give credit to the author.

Useful Documents

MANUALS
I wrote a complete manual of R-CODE, entitled Manual de R-CODE v2.1, only available in Spanish (translators wanted). It covers all aspects of R-CODE for Aibo programming. You can find another R-CODE tutorial in English and Français below, entitled R-Code SDK Tutorial (v1.2). The English one is a limited previous version of the one in Spanish.

R. Téllez, Aibo Quickstart Manual, Internal Report for the Agents Research Lab at the University of Girona, Spain, 2004.

TEACHING MATERIALS
Teaching material for the Second Summer Course on Aibo programming (July 2005)

Introduction to the Aibo OPEN-R programming environment
Presentation Slides
The neural leg controller using OPEN-R

Teaching material for the Mediterranean University Campus Summer Course on Aibo programming (Sep 2004)

R-Code SDK Tutorial (v1.2). Traduction au Français par TME, disponible ici
OPEN-R Essentials
Presentation slides

Useful Code

OPENR CODE

    == Neural network control of Aibo joints

    This code is a beautiful example of OPEN-R program for all those people interested in learning OPEN-R. The code controls one leg of Aibo (all three joints) by means of three different neural networks. Neural networks make the all three joints perform an oscillation. The weights of the neural nets where previously obtained on the Webots simulator, and are included in the form of files.
    This code includes a small library containing the definition of the neural net object and shows how an OPEN-R program can be linked with external objects. It also shows how to access data stored in the memory stick.

    The neural code here.
    More information in this guide I published in 2005.

    == Webots server for Aibo

    I have contributed to the creation of the Webots server for Aibo. This is an OPENR program that runs on Aibo acting as a server and allows the Webots simulator to interact with the real robot as well as cross-compile code from the simulator to the real robot. The source code has been released to the community for free by the owner of the simulator (Cyberbotics).

    The source code is available here and you will find more information about it here.

URBI CODE
    == Neural control of Aibo walking using libUrbi

    This is C++ code that uses the libUrbi library to make Aibo walk. The walking of the robot is achieved by means of neural networks that are implemented in C++ code. Weights for the neural networks were obtained previously using Webots, but are also provided in the code. There are two different versions of this program, the asynchronous one and the synchronous one. More information about those controllers can be found at my evolutionary gaits page.

    Synchronous libUrbi code here.
    Asynchronous libUrbi code here.

    == Aibo dance scripts

    The following Urbi scripts where created for a Aibo dancing demo. These are two scripts that make two Aibo robots dance the Daft Punk song Technology. The scripts do not include the song because of legal issues. You can watch the scripts in action here.

    Urbi script for the left Aibo here.
    Urbi script for the right Aibo here.

    == Aibo follows a person at close distance

    Very simple Urbi script that I call Urage which makes Aibo follow a person dressing in red an stay at a given distance of him. When the person stops, Aibo walks to him until it reaches the given distance. The person can turn and Aibo will also do it.

    Urage script here.

WEBOTS CODE

Not available yet


Useful Links

    OPEN-R tutorial by Baille and Serra at ENSTA
    Tekkotsu open source programming API for OPEN-R
    URBI scripted language to remotely control Aibo
    Pyro allows the use of Python language to control Aibo
    Webots software robot simulator including Aibo

Web page by R. Téllez using rubric css by Hadley Wickham
Don't undertake a project unless it is manifestly important and nearly impossible (Edwin Land)