Locally-Optimal Navigation in Multiply-Connected Environments without Geometric Maps

Benjamin Tovar, Steven M Lavalle, Rafael Murrieta

Research output: Contribution to conferencePaper

Abstract

In this paper we present an algorithm to build a sensor-based, dynamic data structure useful for robot navigation in an unknown, multiply-connected planar environment. This data structure offers a robust framework for robot navigation, avoiding the need of a complete geometric map or explicit localization, by building a minimal representation based entirely on critical events in online sensor measurements made by the robot. There are two sensing requirements for the robot: it must detect when it is close to the walls, to perform wall-following reliably, and it must be able to detect discontinuities in depth information. It is also assumed that the robot is able to drop, detect and recover a marker. The navigation paths generated are optimal up to the homotopy class to which the paths belong, even though no distance information is measured.

Original languageEnglish (US)
Pages3491-3497
Number of pages7
StatePublished - Dec 26 2003
Event2003 IEEE/RSJ International Conference on Intelligent Robots and Systems - Las Vegas, NV, United States
Duration: Oct 27 2003Oct 31 2003

Other

Other2003 IEEE/RSJ International Conference on Intelligent Robots and Systems
CountryUnited States
CityLas Vegas, NV
Period10/27/0310/31/03

    Fingerprint

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Software
  • Computer Vision and Pattern Recognition
  • Computer Science Applications

Cite this

Tovar, B., Lavalle, S. M., & Murrieta, R. (2003). Locally-Optimal Navigation in Multiply-Connected Environments without Geometric Maps. 3491-3497. Paper presented at 2003 IEEE/RSJ International Conference on Intelligent Robots and Systems, Las Vegas, NV, United States.