Skip to main content
main-content

Über dieses Buch

This monograph introduces a unifying framework for mapping, planning and exploration with mobile robots considering uncertainty, linking such problems with a common SLAM approach, adopting Pose SLAM as the basic state estimation machinery. Pose SLAM is the variant of SLAM where only the robot trajectory is estimated and where landmarks are used to produce relative motion measurements between robot poses. With regards to extending the original Pose SLAM formulation, this monograph covers the study of such measurements when they are obtained with stereo cameras, develops the appropriate noise propagation models for such case, extends the Pose SLAM formulation to SE(3), introduces information-theoretic loop closure tests, and presents a technique to compute traversability maps from the 3D volumetric maps obtained with Pose SLAM. A relevant topic covered in this monograph is the introduction of a novel path planning approach that exploits the modeled uncertainties in Pose SLAM to search for the path in the pose graph that allows the robot to navigate to a given goal with the least probability of becoming lost. Another relevant topic is the introduction of an autonomous exploration method that selects the appropriate actions to drive the robot so as to maximize coverage, while minimizing localization and map uncertainties. This monograph is appropriate for readers interested in an information-theoretic unified perspective to the SLAM, path planning and exploration problems, and is a reference book for people who work in mobile robotics research in general.

Inhaltsverzeichnis

Frontmatter

Chapter 1. Introduction

Abstract
Simultaneous localization and mapping (SLAM) is the process where a mobile robot builds a map of an unknown environment while at the same time being localized relative to this map.
Rafael Valencia, Juan Andrade-Cetto

Chapter 2. SLAM Front-End

Abstract
In this Chapter we discuss our choice of front-end for SLAM, the part in charge of processing the sensor information to generate the observations that will be fed to the estimation machinery.
Rafael Valencia, Juan Andrade-Cetto

Chapter 3. SLAM Back-End

Abstract
The SLAM problem has been traditionally addressed as a state estimation problem in which perception and motion uncertainties are coupled.
Rafael Valencia, Juan Andrade-Cetto

Chapter 4. Path Planning in Belief Space with Pose SLAM

Abstract
The probabilistic belief networks that result from standard feature-based simultaneous localization and map building methods cannot be directly used to plan trajectories.
Rafael Valencia, Juan Andrade-Cetto

Chapter 5. Active Pose SLAM

Abstract
This Chapter presents an active exploration strategy that complements Pose SLAM and the path planning approach shown in Chap. 4.
Rafael Valencia, Juan Andrade-Cetto

Chapter 6. Conclusions

Abstract
The work presented in this book constitutes a step towards an integrated framework for mapping, planning and exploration for autonomous mobile robots.
Rafael Valencia, Juan Andrade-Cetto
Weitere Informationen