This paper presents a novel method for autonomous robotic navigation and mapping of large-scale spaces with minimal sensing. The proposed algorithm constructs a graph-based map that encodes the relative location of landmarks in the environment. Uncertainty in these locations is captured by imposing qualitative constraints on the relationships between landmarks observed by the robot. These relationships are represented in terms of the relative geometrical layout of landmark triplets. A novel measurement method based on camera imagery is presented that extends previous work from the field of qualitative spatial reasoning. Measurements are fused into the map using a deterministic approach based on iterative graph updates. The generation of these maps does not depend on estimates of robot egomotion, and it is consequently suitable for high-slip environments. Algorithm performance is evaluated using Monte Carlo simulations, and results are presented for an experiment using data gathered in the Jet Propulsion Laboratory, California Institute of Technology MarsYard.
Qualitative Relational Mapping for Mobile Robots with Minimal Sensing
Journal of Aerospace Information Systems ; 11 , 8 ; 497-511
2014-08-01
Conference paper , Article (Journal)
Electronic Resource
English
Qualitative Relational Mapping for Autonomous Robotics
AIAA | 2012
|2012-2439 Qualitative Relational Mapping for Robotic Navigation
British Library Conference Proceedings | 2012
|Qualitative Relational Mapping for Planetary Rover Exploration (AIAA 2013-5037)
British Library Conference Proceedings | 2013
|Gaussian processes autonomous mapping and exploration for range-sensing mobile robots
British Library Online Contents | 2018
|