PeRL STUDIES AUTONOMOUS NAVIGATION & MAPPING FOR MOBILE ROBOTS IN A PRIORI UNKNOWN ENVIRONMENTS.

At a Glance

Synopsis

Browse Publications by Ryan Eustice and the rest of the PeRL Team.

Browse by year

2017, 2016, 2015, 2014, 2013, 2012, 2011, 2010, 2009, 2008, 2007, 2006, 2005, 2004, 2003, 2002, 2000

Theses

Next-best-view visual SLAM for bounded-error area coverage

Summary


Ayoung Kim and Ryan M. Eustice, Next-best-view visual SLAM for bounded-error area coverage. In IROS Workshop on Active Semantic Perception, Vilamoura, Portugal, October 2012.

Abstract

Navigating an unexplored environment using simultaneous localization and mapping (SLAM) requires that the robot's trajectory include revisit actions in order to produce loop-closure constraints; however, efficient area coverage requires that the robot's trajectory be minimally redundant in its path. This paper reports on a next-best-view SLAM algorithm that balances the trade-off between exploration and revisiting actions in order to simultaneously achieve efficient target area coverage and bounded-error navigation performance. Since area coverage efficiency and bounded localization performance represent competing objectives, the proposed algorithm computes the next-best control action required for localization and area coverage performance. The proposed algorithm, called perception-driven navigation (PDN), represents an integrated navigation solution to the robotic area coverage problem whereupon visual SLAM perception uncertainty is explicitly accounted for. Results are shown for simulated monocular visual SLAM trajectories representative of the type of area coverage problem encountered in autonomous underwater ship hull inspection.

Bibtex entry

@CONFERENCE { akim-2012a,
    AUTHOR = { Ayoung Kim and Ryan M. Eustice },
    TITLE = { Next-best-view visual {SLAM} for bounded-error area coverage },
    BOOKTITLE = { IROS Workshop on Active Semantic Perception },
    YEAR = { 2012 },
    MONTH = { October },
    ADDRESS = { Vilamoura, Portugal },
}