Navigation à partir d'une mémoire d'images

Anthony Remazeilles 1, 2
1 Lagadic - Visual servoing in robotics, computer vision, and augmented reality
CRISAM - Inria Sophia Antipolis - Méditerranée , Inria Rennes – Bretagne Atlantique , IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE
2 TEXMEX - Multimedia content-based indexing
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, Inria Rennes – Bretagne Atlantique
Abstract : This work deals with autonomous robotic navigation. A robotic system that moves on its own must be able to localise itself in the environment, so as to define a path to follow, and also to control its motion in order to reach a desired position. A vision based-approach is chosen, by onsidering robotic systems with an on-board camera. The navigation space is represented in a topological way by an image database acquired during a learning step. Thus, the robot localisation is directly related to a search in the database of the views that are the closest in terms of content from the one acquired by the camera. This search is performed using image retrieval methods. A sequence of overlapping views is defined by the use of topological graphs associated with the database. These views visually describe the zone the robot must traverse to reach the desired position. Robot motion is computed on-line for each image acquired by the camera. The proposed method is based on potential field and visual servoing principles. The control law constrains the robot to move inside an area where the observability of landmarks is considered acceptable. Features used during the visual servoing are initially detected in the path of images, and are automatically localised during the motion by using an image transfer method. Experimental results obtained with a 6 d.o.f. robotic arm, and simulations realised show the validity of the navigation method proposed.
Complete list of metadatas

Cited literature [109 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00524505
Contributor : Patrick Gros <>
Submitted on : Friday, October 8, 2010 - 9:28:56 AM
Last modification on : Thursday, November 15, 2018 - 11:57:20 AM
Long-term archiving on : Monday, January 10, 2011 - 11:38:34 AM

Identifiers

  • HAL Id : tel-00524505, version 1

Citation

Anthony Remazeilles. Navigation à partir d'une mémoire d'images. Interface homme-machine [cs.HC]. Université Rennes 1, 2004. Français. ⟨tel-00524505⟩

Share

Metrics

Record views

777

Files downloads

729