Long-term Ground Robot Localization Architecture for Mixed Indoor-Outdoor Scenarios
Javier Pérez-Lara, Fernando Caballero, and Luis Merino. Long-term Ground Robot Localization Architecture for Mixed Indoor-Outdoor Scenarios. In Proceedings of the International Symposium on Robotics, ISR, 2014.
Download
Abstract
This paper summarizes the validation and experimental results of an architecture for six degree-of-freedom robot localization developed in the framework of the EC funded project FROG (FP7-ICT-2011.2.1). Two main localization issues are considered; one is accuracy, required by the Augmented Reality application, and the second is robustness, in order to achieve long-term autonomy of the robot. The experiments were carried out mainly at the Lisbon Zoo (Portugal), a low GPS visibility area with more than 40,000 square meters and non-planar routes as long as 1 kilometer. The approach considers an offline SLAM and multi-sensor data fusion for map building, and a Rao-Blackwellized filter for online robot localization based on previously computed map. The approach also considers localization failures and provides a method for robot re-localization based on visual place recognition.
BibTeX Entry
@INPROCEEDINGS{isr14, author = {Javier P\'{e}rez-Lara and Fernando Caballero and Luis Merino}, title = {Long-term Ground Robot Localization Architecture for Mixed Indoor-Outdoor Scenarios}, booktitle = {Proceedings of the International Symposium on Robotics, ISR}, year = {2014}, abstract={This paper summarizes the validation and experimental results of an architecture for six degree-of-freedom robot localization developed in the framework of the EC funded project FROG (FP7-ICT-2011.2.1). Two main localization issues are considered; one is accuracy, required by the Augmented Reality application, and the second is robustness, in order to achieve long-term autonomy of the robot. The experiments were carried out mainly at the Lisbon Zoo (Portugal), a low GPS visibility area with more than 40,000 square meters and non-planar routes as long as 1 kilometer. The approach considers an offline SLAM and multi-sensor data fusion for map building, and a Rao-Blackwellized filter for online robot localization based on previously computed map. The approach also considers localization failures and provides a method for robot re-localization based on visual place recognition. }, }
Generated by bib2html.pl (written by Patrick Riley ) on Mon Jan 09, 2023 18:40:36