Project

Back to overview

A Modeling and Control Methodological Framework for Self-Assembling Floating Miniature Robotic Systems

English title A Modeling and Control Methodological Framework for Self-Assembling Floating Miniature Robotic Systems
Applicant Martinoli Alcherio
Number 157191
Funding scheme Project funding (Div. I-III)
Research institution Laboratoire de systèmes et algorithmes intelligents distribués EPFL - ENAC - IIE - DISAL
Institution of higher education EPF Lausanne - EPFL
Main discipline Other disciplines of Engineering Sciences
Start/End 01.11.2014 - 31.10.2017
Approved amount 121'092.00
Show all

All Disciplines (3)

Discipline
Other disciplines of Engineering Sciences
Electrical Engineering
Information Technology

Keywords (5)

distributed stochastic control; swarm robotics; multi-level modeling ; miniature embedded systems; self-assembly

Lay Summary (French)

Lead
Au cours des dernières années, les avancées technologiques ont orientée la communauté robotique vers le domaine de la haute miniaturisation. Les robots ultra-petits peuvent accéder à des environnements qui sont hors de la portée des plates-formes robotiques typiques. Cependant, la miniaturisation est au prix d’une capacité computationnel, sensoriel, d’actuation et de communication minimaliste. Ces restrictions sévères créent la nécessité d'une solution collaborative des tâches en misant sur les comportements collectifs auto-organisés.
Lay summary

Ce projet représente la deuxième phase d'un effort de recherche de quatre ans visant à efficacement concevoir, modéliser et commander des systèmes robotiques auto-organisés. Ce projet entend tirer parti de notre plate-forme expérimentale développée pendant la première phase et construite autour d’un dispositif flottant intelligent à taille centimétrique, appelé Lily. Les Lilys ont été conçus comme émulateurs de futurs micro-dispositifs dotés de fonctionnalités similaires mais d’ordres de magnitudes plus petites.

Cette deuxième phase se focalise sur le développement de systèmes de contrôles stochastiques et distribués étroitement intégrés dans un cadre de modélisation multi-niveaux, qui doit permettre la prédiction et l'optimisation de la dynamique de notre plate-forme robotique distribuée en fonction d'une grande variété de paramètres de conception et de contrôle liant fortement les deux phases de recherche.

En résumé, ce projet vise à: (i) tirer parti de notre plate-forme expérimentale robotique distribuée; (ii) insérer la plate-forme dans un seul cadre de modélisation multi-niveaux; et (iii) d'explorer de façon distribué, stochastique et originale les approches de contrôle.

Direct link to Lay Summary Last update: 29.10.2014

Responsible applicant and co-applicants

Employees

Publications

Publication
Probabilistic Modeling of Programmable Stochastic Self-Assembly of Robotic Modules
Haghighat Bahar, Thandiackal Robin, Mordig Maximilian, Martinoli Alcherio (2017), Probabilistic Modeling of Programmable Stochastic Self-Assembly of Robotic Modules, in IEEE/RSJ International Conference on Intelligent Robots and Systems, Vancouver.
Automatic synthesis of rulesets for programmable stochastic self-assembly of rotationally symmetric robotic modules
Haghighat Bahar, Martinoli Alcherio (2017), Automatic synthesis of rulesets for programmable stochastic self-assembly of rotationally symmetric robotic modules, in Swarm Intelligence, 11(3-4), 243-270.
Characterization and Validation of a Novel Robotic System for Fluid-Mediated Programmable Stochastic Self-Assem
Haghighat Bahar, Martinoli Alcherio (2016), Characterization and Validation of a Novel Robotic System for Fluid-Mediated Programmable Stochastic Self-Assem, in IEEE/RSJ International Conference on Intelligent Robots and Systems, DaejeonIEEE, Daejeon.
Synthesizing rulesets for programmable robotic self-assembly: A case study using floating miniaturized robots
Haghighat Bahar, Platerrier Brice, Waegeli Loic, Martinoli Alcherio (2016), Synthesizing rulesets for programmable robotic self-assembly: A case study using floating miniaturized robots, in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and , Bruxelles 9882 LNCS, 197-209, Springer International Publishing, Switzerland 9882 LNCS, 197-209.
Fluid-mediated stochastic self-assembly at centimetric and sub-millimetric scales: Design, modeling, and control
Haghighat Bahar, Mastrangeli Massimo, Mermoud Grégory, Schill Felix, Martinoli Alcherio (2016), Fluid-mediated stochastic self-assembly at centimetric and sub-millimetric scales: Design, modeling, and control, in Micromachines, 7(8), 138.
A Rule Synthesis Algorithm for Programmable Stochastic Self-Assembly of Robotic Modules
Haghighat Bahar, Martinoli Alcherio, A Rule Synthesis Algorithm for Programmable Stochastic Self-Assembly of Robotic Modules, in The 13th International Symposium on Distributed Autonomous Robotic Systems, London, UK.

Collaboration

Group / person Country
Types of collaboration
Prof. Ani Hsieh (Drexel University) United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results
Prof. Juergen Brugger (EPFL) Switzerland (Europe)
- in-depth/constructive exchanges on approaches, methods or results

Scientific events

Active participation

Title Type of contribution Title of article or contribution Date Place Persons involved
IEEE/RSJ Int. Conf. on Intelligent Robots and Systems Talk given at a conference Probabilistic Modeling of Programmable Stochastic Self-Assembly of Robotic Modules 24.09.2017 Vancouver, Canada Haghighat Bahar;
13th Int. Symp. on Distributed Autonomous Robotic Systems Talk given at a conference A Rule Synthesis Algorithm for Programmable Stochastic Self-Assembly of Robotic Modules 06.11.2016 London, Great Britain and Northern Ireland Haghighat Bahar;
IEEE/RSJ Int. Conf. on Intelligent Robots and Systems Talk given at a conference Characterization and Validation of a Novel Robotic System for Fluid-Mediated Programmable Stochastic Self-Assembly 09.10.2016 Daejeon, Korean Republic (South Korea) Haghighat Bahar;
10th Int. Conf. on Swarm Intelligence (ANTS 2016) Talk given at a conference Synthesizing Rulesets for Programmable Robotic Self-Assembly: A Case Study using Floating Miniaturized Robots 07.09.2016 Bruxelles, Belgium Haghighat Bahar;


Awards

Title Year
3rd place prize of MT180 competition at EPFL 2017
Public vote prize of Pitch Your Impact competition organized by ENAC school at EPFL 2017
Swiss National Science Foundation Early Postdoc.Mobility fellowship 2017

Associated projects

Number Title Start Funding scheme
68647 Design, Control, and Optimization of Swarm-Intelligent, Real-Time, Embedded Systems 01.08.2003 SNSF Professorships
137838 Design, Modeling, and Control Methodologies for Self-Assembling Floating Miniature Robotic Systems 01.11.2012 Project funding (Div. I-III)
116913 Design, Control, and Optimization of Swarm-Intelligent, Real-Time, Embedded Systems 01.08.2007 SNSF Professorships
137838 Design, Modeling, and Control Methodologies for Self-Assembling Floating Miniature Robotic Systems 01.11.2012 Project funding (Div. I-III)
113795 Trajectory Analysis and Behavioral Identification in Mobile Robotic Systems 01.01.2007 Project funding (Div. I-III)
105565 Analyse des trajectoires et identification comportementale pour des systèmes robotique collectifs 01.01.2005 Project funding (Div. I-III)

Abstract

In recent years the technological breakthroughs in the field of µ- and nano-engineering are steering the robotics community towards the realm of high miniaturization. Ultra-small robots can in principle access environments that are beyond the reach of typical robotic platforms, and thus hold great promises in a large variety of disciplines, including biomedical engineering, pervasive information technology, and environmental engineering. However, miniaturization comes at a price: minimalist computation, sensing, actuation, and communication capabilities. These severe restrictions create the need for a collaborative approach towards the solution of non-trivial tasks by leveraging self-organized collective behaviors. In swarm robotics a large body of research has been dedicated to study and design of distributed, stochastic control strategies that are scalable both in terms of swarm size (i.e., they can be implemented in swarms of thousands of robots or more) and individual node size (i.e., they can be implemented using miniature robots with very limited capabilities). Within the broader perspective of investigating how one can transpose and adapt such control strategies from the centimeter-scale down to the (sub-)millimeter-scale, this proposed project represents the second phase of an overall four year research effort aiming at efficiently engineering, modeling, and controlling self-organized robotic systems. Working at the centimeter scale, where the core of the robotic expertise of our laboratory has been built in recent years, readily allows for the development of new experimental tools (e.g., robotic nodes, monitoring and agitation systems) and methods (e.g., modeling, control, optimization) which, while of considerable importance per se, will represent a solid milestone for future downscaling and adaptation to platforms of much smaller dimensions.At the heart of the project lie the concepts of aggregation and self-assembly. We believe that, regardless of the tremendous promises of MEMS technology, actual long-range communication at sub-millimetric scales will not be feasible before years. Thus, relatively short-range interactions or physical contact will meanwhile remain the most viable options to achieve some form of collaborative information sharing which has the potential to be downscaled to even smaller devices; and physical aggregation is the necessary condition for achieving a collectively coordinated response of the system. A first class of experiments will be concerned with the distributed formation of ordered patterns (i.e., self-assembly). A second, slightly more complex class of experiments will focus on the formation of functional configurations and adaptive structures around specific static or dynamic templates serving as precursors of future, more involved tasks such as object protection, transport, or destruction.This second phase intends leveraging our flexible experimental platform built around centimeter-sized intelligent water floating devices, called Lilies. While Lily robots have been designed using standard mechatronic technology, they have been intrinsically conceived as cm-scale emulators of future micro-devices endowed with similar functionalities but having a two order of magnitude smaller size. Phase I focused on the design, fabrication, and packaging of our fluid-mediated self-assembling experimental platform. Phase II is mainly concerned with the development of distributed, stochastic control schemes tightly integrated with a multi-level modeling framework. Such a modeling framework shall enable prediction and optimization of the dynamics of our distributed robotic platform as a function of a large variety of design and control parameters, therefore strongly linking the two research phases. In summary, this project aims to: (i) leverage our distributed robotic experimental platform as a physical, centimeter-scale test-bed to study stochastic fluidic self-assembly; (ii) capture the resource-constrained platform within a single multi-level modelling framework, allowing for in-depth formal analysis and synthesis of the properties of the self-assembling robotic system; (iii) and explore original distributed, stochastic, and potentially scalable control approaches leveraging a tight coupling with the multi-level modeling framework mentioned above.
-