Přístupnostní navigace
E-application
Search Search Close
Publication detail
JELÍNEK, A. ŽALUD, L.
Original Title
Scan to map fitting for indoor localisation of mobile robots - vectorized approach
Type
conference paper
Language
English
Original Abstract
Localisation with a map and especially a simultaneous localisation and mapping (SLAM) employs researchers in robotics for more than three decades. Large amount of data calls for an effective compression and generalizing technique, which would help to get rid of unnecessary measurements and distinguish between the signal and a noise. We believe, that such technique is vectorization, i.e. conversion of point-like data (originating from distance measurement) into more complex geometrical objects, such as line segments in two dimensional space, or triangle meshes in 3D. To prove a practical usability of our approach in the SLAM related problems, a scan to map registration procedure was devised. It is fast and gives optimal results, if correct correspondences are provided. Though we do not deal with the SLAM in its full complexity yet, the results confirm the vectorization to be a viable way forward and determine further direction of our research.
Keywords
Line segment, registration, slam, feature matching
Authors
JELÍNEK, A.; ŽALUD, L.
Released
20. 4. 2017
Publisher
Masaryk University
Location
Brno
ISBN
978-80-210-8550-3
Book
CEITEC PhD Retreat II, Book of abstracts
Edition
1
Pages from
75
Pages to
Pages count
BibTex
@inproceedings{BUT135006, author="Aleš {Jelínek} and Luděk {Žalud}", title="Scan to map fitting for indoor localisation of mobile robots - vectorized approach", booktitle="CEITEC PhD Retreat II, Book of abstracts", year="2017", series="1", pages="75--75", publisher="Masaryk University", address="Brno", isbn="978-80-210-8550-3" }