EAIA07

Invited Lecturer		Atelier Title
Francesco Mondada		Introduction and Exercices on the E-Puck Robots
Stefano Nolfi		Behaviour as a Complex Adaptive System: Coordinated Behaviours and the Emergence of Simple Communication Forms
Guy Theraulaz + Simon Garnier		The Embodiment and Control of Collective Behaviours in Groups of Robots
Luís Paulo Reis		Definition of High-Level Coordination Techniques for Robotic Teams
Richard Vaughan		Player and Stage for Collective Robotics Research

Introduction and Exercices on the E-Puck Robot

Francesco Mondada (Laboratoire des Systèmes Robotiques - École Polytechnique Fédérale de Lausanne - Swiss)

The E-Puck Robot

This short presentation will introduce a new open-hardware robotic platform called e-puck. The e-puck has been designed as an educational tool but is specially adapted to collective experimentation. This introduction will illustrate the structure of the robot in term of mechanics, electronics and software, giving to the participants the basic knowledge in order to use the robot during the workshops.

Exercices On the E-Puck

During this workshop, each participant (or group of two) will get an e-puck mobile robot and will learn how to program, debug and test it. After some basic programming example, one larger example will allow the participants to implement a collective behavior that will be tested at the end of the workshop. The e-puck will be used also in some of the following workshops.

Return to Top

Behaviour as a Complex Adaptive System: Coordinated Behaviours and the Emergence of Simple Communication Forms

Stefano Nolfi (Laboratory of Autonomuous Robotics and Artificial Life - CNR - Rome - Italy)

During the workshop I will briefly introduce software, that will be distributed to the participants, which allows to run collective evolutionary experiments on wheeled robots. After the introduction, the participants will be encouraged to replicate some pre-tested experiments or to run new experiments on their laptop computers

Return to Top

The Embodiment and Control of Collective Behaviours in Groups of Robots

Guy Theraulaz + Simon Garnier (Centre de Recherches sur la Cognition Animale - Université Paul Sabatier - Toulouse - France)

Abstract unavailable for now

Return to Top

Definition of High-Level Coordination Techniques for Robotic Teams

Luís Paulo Reis (Distributed Artificial Intelligence & Robotics Group - Porto - Portugal)

Definition of high-level coordination techniques for robotic soccer teams (including strategies, tactics, formations, dynamic role assignment, strategic positioning methodologies) using RoboCup soccer simulators and FC Portugal Base Code (European and World Champion of RoboCup 2006 - Simulation 3D League). The Workshop will end with a simulated robosoccer mini-tournament, between teams developed by the students during the practical session.

Interactive Lecture

· Introduction to Soccer Simulation.

· Soccer Server Simulator.

· Examples of RoboCup Soccer Simulator in use.

· FC Portugal Base Code.

· Examples of FC Portugal code in use.

Practical Session - Algorithm Design

· Design of basic collective robosoccer strategies.

· Test of basic collective strategies with different opponents.

· Design of advanced strategies (including tactics, formations, setplays and player types).

· Strategy test and refining for different opponents.

· Advanced coordination algorithms design: Situation Based Strategic Positioning, Dynamic Positioning and Role Exchange, Flexible Setplay Selection and others.

· Design of a simple RoboSoccer team with advanced coordination techniques.

Mini- Tournament

· Simulated RoboSoccer Mini-tournament: Featuring all teams developed by the students.

Notes:

· Attendees are invited to bring laptops with RCSSServer 10.0.7 and RCSSBase 10.0.11 simulator system installed.

· Examples will be coded in C++, using FC Portugal team base code under Linux operating system

Return to Top

Player and Stage for Collective Robotics Research

Richard Vaughan (Autonomy Laboratory - Simon Fraser University - Burnaby)

Player and Stage were designed to be useful to multi-robot systems researchers. This atelier is an opportunity for researchers to meet and explore how to get the most out of these tools. In the first of two sessions, one of the software authors will briefly introduce and describe the system, then lead discussion and give advice on how to conduct large-scale multi-robot experiments, and how to modify the software to meet special requirements. In the second session, researchers will write and test new controllers concurrently in a single Stage simulation.

Session 1
Lecture
- Introduction, architecture, and goals
- Examples of Player and Stage in use
Interactive
   - Writing basic Player clients (brief, if necessary)
   - Methods for writing and running multiple Player clients, homogenous and heterogenous
   - Students discuss their previous experience and requirements
   - Extending Player and Stage
   - Methods for scaling up to large populations
   - Long-duration simulations
   - Automating and visualizing experiments

Session 2
Joint demonstration
   - Delegates write and run new robot controllers concurrently in a single common Stage simulation
Closing remarks
   - P/S development roadmap
   - Ideas for the future of collective robotics

** Notes **
- Attendees are invited to bring laptops with the latest release of Player/Stage installed.
- Client programs examples will be in Python and C. Player and Stage extension examples will be in C and C++.

Return to Top