Exploring Unknown Structured Environments
Abstract
This paper presents perceptual algorithms for corridor navigation, door detection and entry that can be used for exploration of an unknown floor of a building. The algorithms are fairly robust to sensor noise, do not use odometric data, and can be implemented on a low end PC. The primary sensor used is a laser range scanner. Their technique relies on constraint-based search to align a current laser frame with a map built from previous laser scans. The alignment of range scans is computationally expensive and can be unnecessary when navigating in a structured environment.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2001
- Accession Number
- ADA443608
Entities
People
- Alexander Stoytchev
- Jonathan F. Diaz
- Ronald C. Arkin
Organizations
- Georgia Tech