Thèse soutenue

FR
Auteur / Autrice : Brahim Tamadazte
Direction : Nadine Lefort-Piat
Type : Thèse de doctorat
Discipline(s) : Automatique
Date : Soutenance en 2009
Etablissement(s) : Besançon
Partenaire(s) de recherche : autre partenaire : Université de Franche-comté. UFR des sciences et techniques

Mots clés

FR

Mots clés contrôlés

Résumé

FR  |  
EN

The micromanipulation and the microassembly of microcomponents (1 micron metre to 1 mm) in order to produce microsystems are incredibly difficult. At this scale, components are almost not visible to the naked eye and there is a reversal of the importance of the forces: the surface forces (capillarity, Van der Waals, electrostatic. . . ) become predominate compared to volume forces (weight, inertia). It is thus necessary and essential to solve this problem to solve the problems to carry out innovative strategies appropriate to imaging system and vision techniques as well as to manipulation and control strategies. The microassembly involves micromanipulation tasks (positioning, pick, transfer, place. . . ) as weIl as more complex tasks (spatial orientation, insertion. . . ). Our work concerns the use of a vision system (optical microscope) in order to automate simple tasks of manipulating microcomponents and more complex tasks for MEMS assembly. Several control laws have been developed: 2D multi-scale image-based visual servoing for micromanipulation and pose-based visual servoing for 3D MEMS assembly. For both approaches developed, the accuracy and repeatability obtained in the process of handling and assembly are satisfactory. However, above all, the vision system must be calibrated for best performances. To do this, a multiple scale calibration method for calibrating photonic microscopes has been presented and detailed. From the study of the constraints related to the use of such imaging system, 3D vision techniques such as depth-from-focus and pose-from-focus has been developed and integrated to realize to the full-automation of a microassembly workcell.