Skip to Main content Skip to Navigation
Habilitation à diriger des recherches

Vision-based estimation and robot control

Ezio Malis 1
1 AROBAS - Advanced Robotics and Autonomous Systems
CRISAM - Inria Sophia Antipolis - Méditerranée
Abstract : The conception of full autonomous robotic systems is an enduring ambition for humankind. Nowadays theoretical and technological advances have allowed engineers to conceive complex systems that can replace humans in many tedious or dangerous applications. However, in order to considerably enlarge the flexibility and the domain of applications of such autonomous systems we still need to face several scientific problems at the crossroad of many domains like for example artificial intelligence, signal processing and non-linear systems control. Amidst these challenges, the perception of the environment and the interaction of the robotic systems with the environment are fundamental problems in the design of such autonomous systems. Indeed, the performance of an autonomous robot not only depends on the accuracy, duration and reliability of its perception but also on the ability to use the perceived information in automatic control loops to interact safely with the environment despite unavoidable modeling and measurement errors. Thus, automatic environment sensing and modeling, and robust sensor-based robot control are central scientific issues in robotics.

Several exteroceptive sensors are commonly used in robotics: contact sensors (e.g. force sensors, tactile sensors), GPS, sonars, laser telemeters, vision sensors, and even olfactory sensors. However, artificial vision is of particular importance and interest, mainly due to its versatility and extended range of applicability. It can be used both for the perception and modeling of the robot's environment and for the control of the robot itself. In this context, the spectrum of research is vast. In this presentation, I will focus on my research work on vision-based estimation and vision-based control problems. Vision-based estimation refers to the methods and techniques dedicated to the extraction of the information that can be useful not only for the modeling of the environment but also for robot control. In the case of robotic applications, major challenges are to increase the efficiency, the accuracy and the robustness of the estimation from visual data. Vision-based control refers to the methods and techniques dedicated to the use of visual information in automatic control loops. The challenges are to choose the appropriate visual information and to design stable control laws that are robust to modeling and measurement errors.
Document type :
Habilitation à diriger des recherches
Complete list of metadata

Cited literature [129 references]  Display  Hide  Download
Contributor : Ezio Malis <>
Submitted on : Thursday, August 27, 2009 - 1:54:21 PM
Last modification on : Tuesday, July 28, 2020 - 10:18:01 AM
Long-term archiving on: : Tuesday, June 15, 2010 - 10:52:23 PM



  • HAL Id : tel-00411412, version 1



Ezio Malis. Vision-based estimation and robot control. Automatic. Université Nice Sophia Antipolis, 2008. ⟨tel-00411412⟩



Record views


Files downloads