Date of Award

3-24-2016

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Department of Electrical and Computer Engineering

First Advisor

John F. Raquet, PhD.

Abstract

This research addresses the lack of quantitative integrity approaches for vision navigation, relying on the use of image or image rendering techniques. The ability to provide quantifiable integrity is a critical aspect for utilization of vision systems as a viable means of precision navigation. This research describes the development of two unique approaches for determining uncertainty and integrity for a vision based, precision, relative navigation system, and is based on the concept of using a single camera vision system, such as an electro-optical (EO) or infrared imaging (IR) sensor, to monitor for unacceptably large and potentially unsafe relative navigation errors.

The first approach formulates the integrity solution by means of discrete detection methods, for which the systems monitors for conditions when the platform is outside of a defined operational area, thus preventing hazardously misleading information (HMI). The second approach utilizes a generalized Bayesian inference approach, in which a full pdf determination of the estimated navigation state is realized.

These integrity approaches are demonstrated, in the context of an aerial refueling application, to provide extremely high levels (10-6) of navigation integrity. Additionally, various sensitivities analyzes show the robustness of these integrity approaches to various vision sensor effects and sensor trade-offs.

AFIT Designator

AFIT-ENG-DS-16-M-251

DTIC Accession Number

AD1053789

Download

Share

COinS