A5024450337- Robotics and Sensor-Based Localization
- Underwater Vehicles and Communication Systems
- Advanced Vision and Imaging
- Robot Manipulation and Learning
- Image and Object Detection Techniques
- Hand Gesture Recognition Systems
- Soft Robotics and Applications
- Human Pose and Action Recognition
- Modular Robots and Swarm Intelligence
- Underwater Acoustics Research
- Indoor and Outdoor Localization Technologies
- Balance, Gait, and Falls Prevention
- Optical measurement and interference techniques
- Color Science and Applications
- Image Processing Techniques and Applications
- Advanced Image Fusion Techniques
- Maritime Navigation and Safety
- Advanced Image and Video Retrieval Techniques
- Diabetic Foot Ulcer Assessment and Management
- Robotic Locomotion and Control
- Tactile and Sensory Interactions
- Image Enhancement Techniques
- Gait Recognition and Analysis
- Prosthetics and Rehabilitation Robotics
- Vibration and Dynamic Analysis
Université de Toulon
2014-2025
Conception de Systèmes Mécaniques et Robotiques
2020-2022
Institute of Electrical and Electronics Engineers
2022
Gorgias Press (United States)
2022
Laboratoire de Conception et d'Intégration des Systèmes
2020
Research Centre Inria Sophia Antipolis - Méditerranée
2015
Institut national de recherche en informatique et en automatique
2007-2015
Integra (United States)
2008
CEA Paris-Saclay - Etablissement de Fontenay-aux-roses
2007-2008
Commissariat à l'Énergie Atomique et aux Énergies Alternatives
2007-2008
Deformable object manipulation (DOM) is an emerging research problem in robotics. The ability to manipulate deformable objects endows robots with higher autonomy and promises new applications the industrial, services, health-care sectors. However, compared rigid manipulation, of considerably more complex still open problem. Addressing DOM challenges demands breakthroughs almost all aspects robotics: hardware design, sensing, (deformation) modeling, planning, control. In this article, we...
Underwater images are altered by the physical characteristics of medium through which light rays pass before reaching optical sensor. Scattering and wavelength-dependent absorption significantly modify captured colors depending on distance observed elements to image plane. In this paper, we aim recover an scene as if water had no effect propagation. We introduce SUCRe, a novel method that exploits scene's 3D structure for underwater color restoration. By following points in multiple tracking...
This paper presents a method to determine the rough shape of an object. is step in development ldquoone click grasping toolrdquo, tool everyday-life objects for assistant robot dedicated elderly or disabled. The goal quadric that approximates at best unknown object using multi-view measurements. Non-linear optimization techniques are considered achieve this goal. Since multiple views necessary, active vision process order minimize uncertainty on estimated parameters and next view. Finally,...
Visual localization plays an important role in the positioning and navigation of robotics systems within previously visited environments. When visits occur over long periods time, changes environment related to seasons or day-night cycles present a major challenge. Under water, sources variability are due other factors such as water conditions growth marine organisms. Yet, it remains obstacle much less studied one, partly lack data. This paper presents new deep-sea dataset benchmark...
Assistance to disabled people is still a domain in which lot of progress needs be done. The more severe the handicap is, complex are devices, implying increased efforts simplify interactions between man and these devices. In this document we propose solution reduce interaction user robotic arm. system equipped with two cameras. One fixed on top wheelchair (eye-to-hand) other one mounted end effector arm (eye-in-hand). cameras cooperate grasping task "one click". method generic, it does not...
A critical assumption of many multi-view control systems is the initial visibility regions interest from all views. An initialization step proposed for a hybrid eye-in-hand/eye-to-hand grasping system to fulfil this requirement. In paper, object assumed be within eye-to-hand field view, whereas it may not eye-in-hand one. The model unknown and no database used. lies in complex scene with cluttered background. method automatically focus on presented, tested validated multi view robotic system.
This paper proposes a novel visual servoing approach to control the dynamic walk of humanoid robot. Online information is given by an on-board camera. It used drive robot towards specific goal. Our work built upon recent reactive pattern generator that make use model predictive (MPC) modify footsteps, center mass and pressure trajectories track reference velocity. The contribution formulate MPC problem considering feedback. We compare our with scheme decoupling walking gait generation. Such...
Tethers are used to supply power and transfer data for teleoperated robots. They known limit the robot's workspace could have effect of hampering its motion. What if we take advantage tether? In this paper a new visual servoing scheme catenary shaped deformable objects is introduced in order control tether parametric shape by properly moving fixation point. most approaches target object rigid distant from controlled robot. On contrary, paper, attached robot, thus 3D changes while robot...
This paper introduces a visual servoing scheme for humanoid walking. Though most of the existing approaches follow perception-decision-action scheme, we close loop so that control is robust to model error. Our approach based on new reactive pattern generator which modifies, at level, footsteps, center mass and pressure trajectories track reference velocity. And, in this paper, velocity directly given by law. Since, HRP-2 walk induces sway motion disturbs regulation law, introduce law...
This paper presents a new experimental method for the estimation of hydrodynamic parameters underwater vehicles, namely added mass and drag coefficients. The principle is to accurately record movement vehicle in response known time-varying actuating force, find an optimal solution that allows simulating velocity trajectory fits recorded trajectory. obtained using Numerical Integration Fitting method, which, employing error minimization algorithm, successively generates trajectories through...
Some recent visual-based relocalization algorithms rely on deep learning methods to perform camera pose regression from image data. This paper focuses the loss functions that embed error between two poses based regression. Existing are either difficult-to-tune multi-objective or present unstable reprojection errors ground truth 3D scene points and require a two-step training. To deal with these issues, we introduce novel function which is multiplane homography integration. new does not prior...
Tethers are used in underwater operations to supply power and transfer data between robots a surface vessel. The real-time tracking of an cable can be useful make exploration missions teleoperation safer by avoiding collisions entanglement. This paper introduces new method detect estimate the 3D shape parameters umbilical thanks embedded visual depth sensors. tether buoyancy is assumed slightly negative, approximated with catenary curve whose lowest point remain attachment points. estimation...
Predicting the behavior of visual features on image plane over a future time horizon is an important possibility in many different control problems. For example when dealing with occlusions (or other constraints such as joint limits) classical servoing loop, or also more advanced model predictive schemes recently proposed literature. Several possibilities have been to perform initial correction step for then propagating by exploiting measurements currently available camera. But predictions...