Français Anglais
Accueil Annuaire Plan du site
Home > Research results > Dissertations & habilitations
Research results
Ph.D de

Ph.D
Group : Human-Centered Computing

.

Starts on 01/10/2014
Advisor : PIETRIGA, Emmanuel

Funding : contrat doctoral UPS
Affiliation : Université Paris-Sud
Laboratory : LRI - HCC

Defended on 09/10/2017, committee :
Directeur de thèse :
M. Emmanuel PIETRIGA Université Paris-Sud

Rapporteurs :
M. Laurent GRISONI POLYTECH'LILLE et UNIVERSITÉ DE LILLE SCIENCES & TECHNOLOGIES
M. Stéphane CONVERSY ENAC-LII et Université de Toulouse

Examinateurs :
Mme Caroline APPERT Université Paris-Sud
M. Gilles BAILLY Université Pierre et Marie Curie
M. Stéphane HUOT Inria Lille-Nord Europe
Mme Yvonne JANSEN Université Pierre et Marie Curie

Président :
Mme Chantal REYNAUD Université Paris-Sud

Research activities :

Abstract :
This thesis investigates a novel input technique for enriching the gesture vocabulary on a multi-touch surface based on fingers' relative location and passive tokens. The first project, TouchTokens, presents a novel technique for interacting with multi-touch surfaces and tokens. The originality is that these tokens are totally passive (no need for any additional electronic components) and their design features notches that guide users' grasp. The purpose of the notches is to indicate a finger spatial configuration (touch pattern) that is specific to the token. When users hold a token and place it on the surface, touching them simultaneously, the system can recognize the resulting touch patterns with a very high level of accuracy (>95%). This approach works on any touch-sensitive surface and makes it possible to easily build low-cost interfaces that combine no-conductive tangibles and gestural input. This technique supports a new multi-touch input that the system can recognize. However, the interaction is limited to the two-state model of touch interaction as the system only knows the tokens' position and cannot detect tokens that are not touched. In the second project of the thesis, we introduce a laser-cut lattice hinge technique for making the tokens flexible. We then develop a new recognizer that analyzes the micro-movements of the fingers while user are holding and deforming those tokens on the surface. We run three experiments to design and calibrate algorithms for discriminating the three following types of manipulations: (1) when a token is left on the surface rather than taken off it (On/Off); (2) when a token has been bent, and (3) when it is squeezed. Our results show that our algorithms can recognize these three manipulations with an accuracy of: On/Off 90.1%, Bent 91.1% and Squeezed 96,9%. The thesis concludes with the presentation of two tools, TouchTokenBuilder and TouchTokenTracker, for facilitating the development of tailor-made tangibles using a simple direct-manipulation interface. TouchTokenBuilder is a software application that assists interface designers in placing notches on arbitrarily-shaped vector contours for creating conflict-free token sets and warning them about potential conflicts. It outputs two files: a vector-graphics description of all tokens in the set and a numerical description of the geometry of each token. TouchTokenTracker is a software library that takes as input the numerical description produced by TouchTokenBuilder, and enables developers to track the tokens' full geometry (location, orientation and shape) throughout their manipulation on the multi-touch surface.

Ph.D. dissertations & Faculty habilitations
.


POWER-AWARE PROTOCOLS FOR WIRELESS SENSOR NETWORKS
In this thesis, we propose a formal energy model which allows an analytical study of energy consumption, for the first time in the context of population protocols. Population protocols model one special kind of sensor networks where anonymous and uniformly bounded memory sensors move unpredictably and communicate in pairs. To illustrate the power and the usefulness of the proposed energy model, we present formal analyses on time and energy, for the worst and the average cases, for accomplishing the fundamental task of data collection. Two power-aware population protocols, (deterministic) EB-TTFM and (randomized) lazy-TTF, are proposed and studied for two different fairness conditions, respectively. Moreover, to obtain the best parameters in lazy-TTF, we adopt optimization techniques and evaluate the resulting performance by experiments. Then, we continue the study on optimization for the poweraware data collection problem in wireless body area networks. A minmax multi-commodity netow formulation is proposed to optimally route data packets by minimizing the worst power consumption. Then, a variable neighborhood search approach is developed and the numerical results show its efficiency. At last, a stochastic optimization model, namely the chance constrained semidefinite programs, is considered for the realistic decision making problems with random parameters. A novel simulation-based algorithm is proposed with experiments on a real control theory problem. We show that our method allows a less conservative solution, than other approaches, within reasonable time.

USE OF INTRISIC MOTIVATIONS IN REINFORCEMENT LEARNING. APPLICATION TO SECURITY IN MULTI-AGENT SYSTEMS