Terrain-Relative Navigation with Neuro-Inspired Elevation Encoding

Terrain-relative autonomous navigation is a challenging task. In traditional approaches, an elevation map is carried onboard and compared to measurements of the terrain below the vehicle. These methods are computationally expensive, and it is impractical to store high-quality maps of large swaths of...

Full description

Saved in:
Bibliographic Details
Published inIEEE/ION Position Location and Navigation Symposium pp. 851 - 857
Main Authors Michaelson, Kristen, Wang, Felix, Zanetti, Renato
Format Conference Proceeding
LanguageEnglish
Published IEEE 24.04.2023
Subjects
Online AccessGet full text
ISSN2153-3598
DOI10.1109/PLANS53410.2023.10139925

Cover

More Information
Summary:Terrain-relative autonomous navigation is a challenging task. In traditional approaches, an elevation map is carried onboard and compared to measurements of the terrain below the vehicle. These methods are computationally expensive, and it is impractical to store high-quality maps of large swaths of terrain. In this article, we generate position measurements using NeuroGrid, a recently-proposed algorithm for computing position information from terrain elevation measurements. We incorporate NeuroGrid into an inertial navigation scheme using a novel measurement rejection strategy and online covariance computation. Our results show that the NeuroGrid filter provides highly accurate state information over the course of a long trajectory.
ISSN:2153-3598
DOI:10.1109/PLANS53410.2023.10139925