Skip to Main content Skip to Navigation

Planning of visual servoing tasks for robotics

Alexis Nicolin 1 
1 LAAS-GEPETTO - Équipe Mouvement des Systèmes Anthropomorphes
LAAS - Laboratoire d'analyse et d'architecture des systèmes
Abstract : Robots are increasingly present in the industrial environment. Originally programmed manually by their operators, today’s robots are progressively using paths calculated by programs that simulate the entire work environment. However, modelling or even execution errors require testing periods to ensure that the robot will perform its task correctly.This thesis intends to link the concept of constraint that the motion planner takes into account with the robot’s real-time control software, which manages the correct execution of tasks. The control must take into consideration the data from the robot’s multiple sensors to adapt the planned paths to reality and ensure that its objectives are achieved. This thesis is being carried out within the framework of a CIFRE agreement between the LAAS-CNRS laboratory and the company Airbus Operations. It is also part of ROB4FAM, a common laboratory between thosetwo entities, which aims at studying the future of robotics applied to aeronautical construction.The main contribution of this thesis is the automatic generation of such control software based on the constraints expressed at the planning stage. Until now, this type of correspondence was mostly established manually. Among the different typesof control schemes available, special emphasis has been placed on visual servo control. This is due to the fact that cameras are affordable sensors capable of generating a considerable amount of data. But above all modern computers are now able to process this volume of information efficiently in real-time. Finally, the control software we present in this manuscript uses a strict hierarchy of commands. This means that the robot will seek to ensure the achievement of the first objective, usually related to the preservation of its integrity, before consideringthe following ones. With this architecture, the robot can adapt to minor changes in its objectives while ensuring the safety of the operators, its working environment and its structure itself. The concepts and software developed during this thesis were implemented on a humanoid robot, TALOS, and then on a mobile wheeled robot equipped with amanipulator arm, TIAGo. They were able to accomplish their tasks, despite the displacement of the objects they were going to manipulate between the movement planning stage and the program execution stage.
Document type :
Complete list of metadata
Contributor : ABES STAR :  Contact
Submitted on : Monday, May 23, 2022 - 8:11:11 AM
Last modification on : Monday, July 4, 2022 - 9:11:32 AM


Version validated by the jury (STAR)


  • HAL Id : tel-03659668, version 2


Alexis Nicolin. Planning of visual servoing tasks for robotics. Automatic. INSA de Toulouse, 2022. English. ⟨NNT : 2022ISAT0003⟩. ⟨tel-03659668v2⟩



Record views


Files downloads