Masters Theses

Keywords and Phrases

Autonomous; Controls; Navigation; Robotics; Validation

Abstract

Spacecraft rendezvous and docking are critical mission phases for various applications of spaceflight, including active debris removal (ADR) and in-orbit servicing, assembly, and manufacturing (ISAM). While previous missions utilized humans to perform rendezvous and docking, this style of mission greatly increases safety risks and cannot be implemented on a large scale. Autonomous servicing satellites provide a path towards scalable rendezvous and proximity operations (RPO) because these autonomous agents do not require human intervention. This work presents a lightweight convolutional neural network (CNN) for the navigation system of an agent which analyzes monocular images and predicts the target's position and orientation (pose) without communicating with the target or relying on ground station communication. Such capabilities enable real-time updates of the relative state during RPO entirely onboard the agent. The proposed network architecture highlights the ability of direct regression methods in the satellite imaging domain. To ensure the reliable operation of this neural network, a ground testing facility is also proposed. The laboratory is being designed to evaluate the efficacy of visual navigation systems as well as reinforcement learning-based guidance by simulating satellite motion and the space imaging environment. The current state of the laboratory is discussed, and future improvements to reach Hardware-in-the-Loop (HIL) capabilities are noted.

Advisor(s)

Nandan, Smriti (Paul)

Committee Member(s)

Pernicka, Henry J.
Song, Yun Seong

Department(s)

Mechanical and Aerospace Engineering

Degree Name

M.S. in Aerospace Engineering

Publisher

Missouri University of Science and Technology

Publication Date

Spring 2026

Pagination

ix, 74 pages

Note about bibliography

Includes_bibliographical_references_(pages 67-73)

Rights

© 2026 Logan Banker , All Rights Reserved

Document Type

Thesis - Open Access

File Type

text

Language

English

Thesis Number

T 12583

Download

Share

 
COinS