Relative Navigation Methods for a Multi-Agent, On-Orbit Inspection Mission

On-orbit inspection is often a necessary prerequisite to satellite operations such as servicing and debris removal. In particular, multiple inspectors collaborating to inspect an unknown object can result in a faster and more comprehensive inspection. The contribution of this research is to provide...

Full description

Saved in:
Bibliographic Details
Published in2023 IEEE/ION Position, Location and Navigation Symposium (PLANS) pp. 1303 - 1310
Main Authors Mercier, Mark, Curtis, David
Format Conference Proceeding
LanguageEnglish
Published IEEE 24.04.2023
Subjects
Estimation
Inspection
Limiting
Multi-Agent
Navigation
Relative Navigation
Satellite Inspection
Satellites
Space technology
Sparse matrices
Online AccessGet full text

Cover

More Information
Summary:On-orbit inspection is often a necessary prerequisite to satellite operations such as servicing and debris removal. In particular, multiple inspectors collaborating to inspect an unknown object can result in a faster and more comprehensive inspection. The contribution of this research is to provide a reliable, computationally efficient estimate for the position of each agent relative to a target in a multi-agent inspection. First, an overview of the problem dynamics, background, and methods used to build and solve a factor graph-based estimation technique is provided. Through the implementation of a Sliding Window (SW) filter, a factor graph can be applied to long duration scenarios while reducing processing requirements for estimation. An example scenario of three agents attempting to conduct an inspection while susceptible to a Sun exclusion zone that nulls some measurement information is prescribed. Development continues by comparing results for the developed Sliding Window Factor Graph (SWFG) against two extreme SW sizes of interest: an iterative Kalman filter (SW size = 1) and a full factor graph (SW size ≥ scenario duration). Results reveal the impact of both the Sun exclusion zone and sliding window size on estimation quality and computation time.
ISSN:2153-3598
DOI:10.1109/PLANS53410.2023.10140084