Path tracing on polar depth maps for robot navigation
2012 (English)In: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Springer Berlin/Heidelberg, 2012, 395-404 p.Conference paper (Refereed)
In this paper a Cellular Automata-based (CA) path estimation algorithm suitable for safe robot navigation is presented. The proposed method combines well established 3D vision techniques with CA operations and traces a collision free route from the foot of the robot to the horizon of a scene. Firstly, the depth map of the scene is obtained and, then, a polar transformation is applied. A v-disparity image calculation processing step is applied to the initial depth map separating the ground plane from the obstacles. In the next step, a CA floor field is formed representing all the distances from the robot to the traversable regions in the scene. The target point that the robot should move towards to, is tracked down and an additional CA routine is applied to the floor field revealing a traversable route that the robot should follow to reach its target location.
Place, publisher, year, edition, pages
Springer Berlin/Heidelberg, 2012. 395-404 p.
, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), ISSN 0302-9743 ; 7495 LNCS
cellular automata, floor field, path estimation, polar transformation, v-disparity image
IdentifiersURN: urn:nbn:se:kth:diva-104971DOI: 10.1007/978-3-642-33350-7-41ScopusID: 2-s2.0-84867534992ISBN: 978-364233349-1OAI: oai:DiVA.org:kth-104971DiVA: diva2:567984
10th International Conference on Cellular Automata for Research and Industry, ACRI 2012, 24 September 2012 through 27 September 2012, Santorini Island
QC 201211152012-11-152012-11-142012-11-15Bibliographically approved