Ph.D
Group : Learning and Optimization

Environment-driven Distributed Evolutionary Adaptation for Collective Robotic Systems

Starts on 01/10/2009
Advisor : BREDECHE, Nicolas

Funding : A
Affiliation : Université Paris-Saclay
Laboratory : LRI AO

Defended on 01/03/2013, committee :

Research activities :

Abstract :
This thesis describes some of the work done in the context of the Symbrion project. This project targets the realization of complex tasks which require the cooperation of multiple robots within robotic swarms (at least 100 robots operating together). Among the issues studied by the project are the self-assembly of robots to form complex structures and the self-organization of large number of robots toward the realization of a common task. Subjects of interests are thus modular self-adaptive robots with both strong coordination properties, and swarm-level cooperation.

The challenge faced by this project is that robots are used in open environments which remain unknown until their deployment. Since operational conditions can’t be predicted beforehand, on-line learning algorithms must be used to design behaviors. In the use of large groups of robots, multiple considerations have to be taken into account such as reduced communication abilities, small memory storage, small computational power. In this context, on-line learning algorithms must be distributed among robots as central control is not an option.

Within this context, the thesis address the problem of maintaining swarm integrity (i.e. to maximize the number of "active" robots). We introduce and define the problem of environment-driven distributed evolutionary adaptation, and present an algorithm to solve this problem. This algorithm has been validated both in simulation and with real robots, and several case studies have been investigated. In particular, we studied the dynamics of the algorithm under various environmental conditions. In particular: evolutionary dynamics and convergence to stable behaviors, robustness to environmental changes and evolution of altruistic behavior in adversarial environment.

CAUSAL LEARNING FOR DIAGNOSTIC SUPPORT


CAUSAL UNCERTAINTY QUANTIFICATION UNDER PARTIAL KNOWLEDGE AND LOW DATA REGIMES


MICRO VISUALIZATIONS: DESIGN AND ANALYSIS OF VISUALIZATIONS FOR SMALL DISPLAY SPACES
The topic of this habilitation is the study of very small data visualizations, micro visualizations, in display contexts that can only dedicate minimal rendering space for data representations. For several years, together with my collaborators, I have been studying human perception, interaction, and analysis with micro visualizations in multiple contexts. In this document I bring together three of my research streams related to micro visualizations: data glyphs, where my joint research focused on studying the perception of small-multiple micro visualizations, word-scale visualizations, where my joint research focused on small visualizations embedded in text-documents, and small mobile data visualizations for smartwatches or fitness trackers. I consider these types of small visualizations together under the umbrella term ``micro visualizations.'' Micro visualizations are useful in multiple visualization contexts and I have been working towards a better understanding of the complexities involved in designing and using micro visualizations. Here, I define the term micro visualization, summarize my own and other past research and design guidelines and outline several design spaces for different types of micro visualizations based on some of the work I was involved in since my PhD.