Degree Type
Dissertation
Date of Award
2016
Degree Name
Doctor of Philosophy
Department
Electrical and Computer Engineering
Major
Computer Engineering
First Advisor
Arun K. Somani
Abstract
GPS is a critical sensor for Unmanned Aircraft Systems (UASs) navigation due to its accu
racy, global coverage, and small hardware footprint. However, GPS is subject to interruption or denial due to signal blockage or RF interference. In such a case, position, velocity and altitude (PVA) performance from other inertial and air data sensor is not sufficient for UAS platforms to continue their primary missions, especially for small UASs.
Recently, image-based navigation has been developed to address GPS outages for UASs, since most of these platforms already include a camera as standard equipage. This thesis develops a novel, automated UAS navigation augmentation scheme, which utilizes publicly available open source geo-referenced vector map data, in conjunction with real-time optical imagery from on-board monocular camera to augment UAS navigation in GPS denied terrain environments. The main idea is to analyze and use terrain drainage patterns for GPS-denied navigation of small UASs, such as ScanEagle, utilizing a down-looking fixed monocular imager. We leverage the analogy between terrain drainage patterns and human fingerprints, to match local drainage patterns to GPU (Graphics Processing Unit) rendered parallax occlusion maps of geo-registered radar returns (GRRR). The matching occurs in real-time. GRRR is assumed to be loaded on-board the aircraft pre-mission, so as not to require a scanning aperture radar during the mission. Once a successful match is made, using a known lens model a final PVA
solution can be obtained from the extrinsic matrix of the camera [1]. Our approach allows
extension of UAS missions to GPS denied terrain areas, with no assumption of human-made geographic objects.
We study the influence of granularity of terrain drainage patterns on performance of our
minutiae-based terrain matching approach. Based on experimental observations, we conclude that our approach delivers a satisfactory performance. We identify the conditions to achieve the desired performance for the input images based on UAS flight altitudes.
Copyright Owner
Teng Wang
Copyright Date
2016
Language
en
File Format
application/pdf
File Size
112 pages
Recommended Citation
Wang, Teng, "Augmented UAS navigation in GPS denied terrain environments using synthetic vision" (2016). Graduate Theses and Dissertations. 15835.
https://lib.dr.iastate.edu/etd/15835
