« Swarm intelligence » et robotique en essaim
Marco Dorigo

Interview de Marco Dorigo dans « les Chercheurs d’Avenir. »

La vidéo a remporté la compétition qui a mis l’équipe de Marco Dorigo face à des équipes prestigieuses provenant des institutions suivantes: NASA, MIT, Carnegie Mellon University, George Mason University, Technische Universitaet Munchen, University of California, Georgia Institute of Technology.

Dernières nouvelles:

  • Nouveau paradigme de coordination des systèmes multi-robots (Nature Communications, sept. 2017)

Il s’agit d’une avancée très significative dans le contrôle de systèmes multi-robots. En effet, auparavant, au sein d’un système multi-robots, chaque robot était indépendant et le paradigme de coordination était l’auto-organisation. Ceci implique une coordination stricte entre les robots, comme par exemple, dans le robot metamorphique (essaim composé de robots indépendants).  

Dans ce nouvel article, le Prof. Marco Dorigo prouve qu’il est possible d’obtenir cette coordination stricte via le concept de “mergeable nervous system”. 

Il a en effet montré qu’il est désormais possible à des robots de transférer toutes leurs informations/expériences et d’abandonner leur indépendance (merging) à un robot donné qui prendrait alors le contrôle de tout l’essaim.

https://www.nature.com/articles/s41467-017-00109-2

  • Le Prof. Marco Dorigo reçoit le IEEE Evolutionary Computation Pioneer Award 2016 pour ses recherches en intelligence artificielle.
  • Le Prof. Marco Dorigo reçoit le 21 avril 2015 les Insignes de Docteur Honoris Causa de l’Université de Pretoria.
  • Le Prof. Marco Dorigo reçoit le 2015 IEEE Frank Rosenblatt Award « pour ses contributions aux fondations de la swarm intelligence »
  • Le projet coordonné par Marco Dorigo « Swarmanoid » remporte le prix vidéo 2011 de l’ Association for Advancement of Artificial Intelligence et le Prix Wernaers 2012 pour la recherche et la diffusion des connaissances.
  • Marco Dorigo bénéficie de l’ ERC Advanced Grant 2010.

Biographie de Marco Dorigo

dorigo

Marco Dorigo, né en 1961. Docteur en Systems and Information electronic Engineering de l’Ecole Polytechnique de Milan (1992), Agrégé de l’Enseignement Supérieur de l’ULB (1995).

En savoir plus sur Marco Dorigo

Directeur de recherche au FNRS, Co-Directeur de l’Institut de Recherches Interdisciplinaires et de Développements en Intelligence Artificielle (IRIDIA, ULB), Rédacteur en chef de la revue « Swarm Intelligence » et éditeur associé et membre du comité éditorial de nombreuses revues scientifiques internationales. Inventeur de la méthode d’optimisation « ant colony optimization » basée sur le comportement de recherche de nourriture des fourmis et co-fondateur du domaine de recherche portant sur la « swarm intelligence ». Il a été distingué par de nombreux prix et financements prestigieux dont notamment : le Association for the Advancement of Artificial Intelligence video award (2011, 2007), l’ERC Advanced Grant (2010), le Cajastur International Prize (2007), le prix quinquennal du FNRS (2005), le Marie Curie Excellence Award (2003), le prix Marco Somalvico pour l’intelligence artificielle (1996).

Le projet

L'équipe de recherche du laboratoire du Dr. Marco Dorigo.

L’équipe de recherche du laboratoire du Dr. Marco Dorigo.

1. Intelligence artificielle et ingénierie robotique en essaim.

Le domaine de recherche du Pr. Marco Dorigo est l’ingénierie robotique en essaim, également appelée « swarm bot ».

Autrement dit la conception et la mise en oeuvre de système d’intelligence artificielle pour résoudre de multiples problèmes pratiques. Je m’inspire des modèles observés par les biologistes d’après l’étude des comportements des animaux sociaux, en particulier des fourmis pour créer des algorythmes qui vont permettre de programmer les robots.

2. Le « swarm bot » ou une collectivité de robots auto-organisés

Qui dit « swarm bot » dit collectivité de robots.

L’idée est de déveloper une légion de petits robots qui peuvent s’auto-assembler, interagir entre eux afin de se répartir une tâche à accomplir. Chacun fonctionnant à la fois en groupe et indépendamment les uns des autres. L’avantage réside dans le fait qu’ensemble, ils sont plus efficaces. Si l’un d’entre eux tombe en panne, les autres restent fonctionnels. Un aspect non négligeable lors d’une mission spatiale par exemple. En groupe, ils sont capables de franchir un obstacle, de se déplacer en terrain accidenté ou de porter une charge plus lourde.

Robots indépendants capables de s'assembler afin d'accomplir des tâches complexes

Robots indépendants capables de s’assembler afin d’accomplir des tâches complexes

3. Applications visées

A terme, on peut imaginer des applications telles que des interventions de secours lors de l’effondrement d’un immeuble ou l’envoi de robots dans un environnement hostile telle qu’une zone hautement radioactive, là où l’être humain ne peut opérer ou se déplacer.

Une "anarchytecture" accomplissant une mission et qui se désassemblera une fois la tâche achevée

Une « anarchytecture » accomplissant une mission et qui se désassemblera une fois la tâche achevée

4. L’avancée des recherches

Ce volet de la robotique est extrêmement novateur.

Aussi à l’heure actuelle, malgré les avancées significatives provenant notamment grâce à la recherche à l’ULB, il faut encore développer une méthodologie rigoureuse pour que le fruit de ces recherches puisse être appliqué.

Autres informations

Connaître le projet de Marco Dorigo en détails (en anglais) : Le projet en PDF

Quelques liens :

  • Marco Dorigo dans The Economist, le 14 août 2010 : Riders on a swarm © The Economist 2010
  • Marco Dorigo reçoit un prestigieux ERC Advanced Grant 2010 (en anglais) : cliquez ici

Budget (English)

The current expertise available at Prof. Dorigo’s lab covers all the necessities of the project except for the hardware construction of the robots. Accordingly, the main part of the requested funding is for the creation of a professorship in mechatronics oriented to the design and construction of advanced micro-robotic systems.  Additional funding is necessary for setting up a mechatronics lab. The goal is to fund the new chair for a period of five year and to have then the ULB taking over and provide the necessary funding for a permanent professor position. The necessary budget is estimated to 2,8 million of EUR over a period of 5 years, organized as follows:

Personnel

One senior researcher at professor level. Here the goal is to hire a mechatronics expert with proved research record in micro-robotics. The offered position will be “tenure track”, that is, it will become a permanent position after the end of the project.

One technician for the new mechatronics laboratory.

One postdoctoral researcher with a PhD in autonomous robotics and mechatronics.

One PhD student with a master in robotics.

Consommables

This concerns the material necessary for the construction of the robots.

Equipments

The set up of the new mecathronics laboratory will be the main investment. Additionally, computational nodes for running simulations will be necessary.

Develop budget table
Needs Operational cost (PER YEAR) Investment
New chair (professorship) in Mecathronics 120000
Salary for technician 65000
Salary for one post-doc 75000
Fellowship for PhD student 34000
Material for building robots 55000
Setup of new laboratory 1000000
Computational nodes 55000
Total over 5 years 1745000 1055000

The Marco Dorigo’s Projet explained in English

Click here to obtain the English version

Anarchytecture: Self-organizing Growing Artifact

The vision

To date, building and construction have always been centralized processes. Central control is implicit in our idea of architecture—the art or science of building. Indeed, the word architecture derives from the Greek arkhitekton (arkhi, chief + tekton, builder), i.e., the chief builder. This project introduces the original idea of artifacts that grow autonomously. They are built as the result of a self-organized process, that is, a process in which there is no leader, no chief builder. Hence the name of the project: Anarchytecture. As customary with names of Greek derivation, the prefix ‘an-’ (privative alpha) expresses negation or absence: Construction without a leader.

The Anarchytecture project will investigate the situation in which humans are taken out of the loop and artifacts build themselves autonomously: they self-build.

The fundamental breakthrough targeted by the Anarchytecture project is the development of techniques for growing and maintaining complex structures using a collection of self-organized autonomous robots as building blocks. In the long term vision that motivates the Anarchytecture project, objects are able to self-build, self-reconfigure, self-repair, and eventually self-dispose and recycle once no longer needed. In other words, we imagine artifacts whose life-cycle is reminiscent of that of living systems. While it is possible to imagine a society in the distant future where all artifacts self-build, relevant applications in the short/medium term might include building in hostile environments where human intervention is difficult or impossible, such as outer space or the ocean floor.

The available expertise

Prof. Dorigo and his team are world leaders in the design of artificial swarm intelligence systems, that is, systems composed of many individuals that coordinate using decentralized control and self-organization. In particular, swarm intelligence focuses on the collective behaviors that result from the local interactions of the individuals with each other and with their environment. Examples of systems studied by swarm intelligence are colonies of ants and termites, schools of fish, flocks of birds, herds of land animals. Prof. Dorigo is reknown for his design and implementation of artificial swarm intelligence systems to tackle optimization and control problems.

In the last 10 years Prof. Dorigo has focused on the design of physical systems composed of many autonomous robots capable of self-assembling into bigger mobile structures capable of performing tasks that are beyond the capabilities of the single robots.  These robots, known as swarm-bots, are the result of some cooperative projects that he coordinated within the “Future and Emerging Technologies” program of the European Commission. Swarm-bots can be seen as an initial, proof of principle of the feasibility of the ideas proposed in the Anarchytecture project.

The necessary tools

Anarchytecture is a multidisciplinary project that will build on the extensive experience of Prof. Dorigo’s group in the design and implementation of distributed, self-organizing swarms of autonomous robots. A multidisciplinary project like Anarchytecture requires the collaboration of scientists and engineers with different background and skills. In particular, it is necessary to be able to exploit the principle of swarm intelligence to propose a high-level design of the anarchytecture distributed system, to design and build the hardware robotic components of the anarchytecture, and to design and implement the control algorithms and communication strategies that let the anarchytecture components self-organize. In addition to these activities, it is necessary to develop a simulation environment where to fast-prototype the hardware ideas before the actual robots are built, and where to test the control algorithms and communication policies before running experiments with the real robots.

The funding requested

The current expertise available at Prof. Dorigo’s lab covers all the necessities of the project except for the hardware construction of the robots. Accordingly, the main part of the requested funding is for the creation of a professorship in mechatronics oriented to the design and construction of advanced micro-robotic systems.  Additional funding is necessary for setting up a mechatronics lab. The goal is to fund the new chair for a period of five year and to have then the ULB taking over and provide the necessary funding for a permanent professor position. The necessary budget is estimated to 2,8 million of EUR over a period of 5 years, organized as follows:

Personnel

One senior researcher at professor level. Here the goal is to hire a mechatronics expert with proved research record in micro-robotics. The offered position will be “tenure track”, that is, it will become a permanent position after the end of the project.

One technician for the new mechatronics laboratory.

One postdoctoral researcher with a PhD in autonomous robotics and mechatronics.

One PhD student with a master in robotics.

Consommables

This concerns the material necessary for the construction of the robots.

Equipments

The set up of the new mecathronics laboratory will be the main investment. Additionally, computational nodes for running simulations will be necessary.

budget_dorigo

Far reaching effects of the project

Making it possible the demonstration of the ideas behind the Anarchytecture project, that is the creation of objects that are able to self-build, self-reconfigure, self-repair, and eventually self-dispose and recycle once no longer needed, will have far reaching effects. It will in fact contribute not only to robotics research, but also to computer science, with the development of novel algorithmic approaches, and to architecture, with the demonstration of a new way of designing and realizing structures that are no longer static, but that can self-build and self-reconfigure to respond to changes in the environment in which they are located and of the function that they are asked to accomplish.

By this project the ULB will become the home of a novel school of thought that reunites computer science, artificial intelligence, robotics and architecture to propose a novel way of designing and implementing self-building artifacts.