CoMapping: Efficient 3D-Map Sharing Methodology for Decentralized cases

Abstract : CoMapping is a framework to efficient manage, share, and merge 3D map data between mobile robots. The main objective of this framework is to implement a Collaborative Mapping for outdoor environments. The framework structure is based on two stages. During the first one, the Pre-Local Mapping stage, each robot constructs a real time pre-local map of its environment using Laser Rangefinder data and low cost GPS information only in certain situations. Afterwards, the second one is the Local Mapping stage where the robots share their pre-local maps and merge them in a decentralized way in order to improve their new maps, renamed now as local maps. An experimental study for the case of decentralized cooperative 3D mapping is presented, where tests were conducted using three intelligent cars equipped with LiDAR and GPS receiver devices in urban outdoor scenarios. We also discuss the performance of the cooperative system in terms of map alignments.
Type de document :
Communication dans un congrès
10th workshop on Planning, Perception and Navigation for Intelligent Vehicles at Int. Conf. on Intelligent Robots and Systems, Oct 2018, Madrid, Spain
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01867720
Contributeur : Olivier Kermorgant <>
Soumis le : mardi 4 septembre 2018 - 15:25:52
Dernière modification le : mercredi 12 septembre 2018 - 01:16:18

Fichier

2018_PPNIV_Contreras_CoMapping...
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01867720, version 1

Citation

Luis Contreras-Samamé, Salvador Domínguez-Quijada, Olivier Kermorgant, Philippe Martinet. CoMapping: Efficient 3D-Map Sharing Methodology for Decentralized cases. 10th workshop on Planning, Perception and Navigation for Intelligent Vehicles at Int. Conf. on Intelligent Robots and Systems, Oct 2018, Madrid, Spain. 〈hal-01867720〉

Partager

Métriques

Consultations de la notice

110

Téléchargements de fichiers

40