Skip to Main content Skip to Navigation

Analyse qualitative de robots

Abstract : This thesis is within the scope of the general problematic of characterizing and classifying systems. More precisely, our contributions are aimed toward robotic applications and, more particularly, at the classification of manipulators. The algorithms and methodologies proposed in this document are based upon several mathematical theories including Singularity Theory and Interval Analysis, which are formalized in the first part of this Phd Thesis report. Classifying systems induces two objectives that are the formalization of the elements commons to systems in the same class as well as their formal and practical detection. Then, this report compiles, for each one in a dedicated chapter, our contributions regarding these two objectives associated with the classification. A general methodology is proposed, with the objective to distinguish several characteristic singular subsets that prove to be generically not empty. This methodology is illustrated for functions between spaces of the same dimension, ranging from 1 to 3. Each application highlights different properties that are admitted by the proposed general methodology. Furthermore, each of these applications leads to the conception of severalalgorithms, based on Interval Analysis, that enforce the proposed methodology. Complementarily to the detection of interest points, we introduce a general methodology to define a combinatorial object encoding the behavior of a function. This methodology is based on the preliminary detection of interest sets whose topologies are invariant for the chosen equivalence. These developments open up to numerous future perspectives that conclude this report.
Document type :
Complete list of metadatas
Contributor : Abes Star :  Contact
Submitted on : Friday, April 6, 2018 - 10:47:11 AM
Last modification on : Friday, May 29, 2020 - 12:40:13 PM


Version validated by the jury (STAR)


  • HAL Id : tel-01760305, version 1


Romain Benoit. Analyse qualitative de robots. Robotique [cs.RO]. Université d'Angers, 2017. Français. ⟨NNT : 2017ANGE0034⟩. ⟨tel-01760305⟩



Record views


Files downloads