Date of Award

Winter 2025

Project Type

Thesis

Program or Major

Mechanical Engineering

Degree Name

Master of Science

First Advisor

May-Win L Thein

Second Advisor

Wayne Smith

Third Advisor

Se Young Yoon

Abstract

The success of future crewed and autonomous space missions hinges on effectively utilizing in-situ resources available in the extraterrestrial environment to sustain missions and habitation. Prior to extraction of resources, such missions will require robust and efficient methods for extraterrestrial resource prospecting. However, traditional global optimization techniques struggle with the complex and non-convex search spaces inherent to mapping and identifying resource deposits on extraterrestrial surfaces.

This research addresses the resource prospecting problem by developing and evaluating novel variants of the Particle Swarm Optimization (PSO). Here, four novel algorithms, their variants, and four combined algorithms are rigorously tested and confirmed. The presented algorithms are designed to balance global exploration with local exploitation, thereby enhancing optimization robustness and convergence reliability.

Each algorithm is confirmed against a varied set of representative fitness functions, including specialized simulated environments based on Lunar and Martian resource landscapes. Simulation results consistently demonstrated the superior performance of the individual and combined algorithms, excluding Voronoi PSO (VPSO). In particular, variations and combinations based upon Transcending PSO (TPSO) and the Continuous PSO (CMPSO) variants showed the strongest performance.

Historical experimental data sets are used in this work to demonstrate the fidelity of modern, hardware-in-the-loop robotic simulations as an experimental platform. Each algorithm presented here is tested through these experimental simulations and compared to the computational simulation results. Through these tests parity is established between the computational and experimental findings such that the computational findings are validated.

The modular PSO methodologies established here are highly effective at navigating complex search topographies. The proposed algorithms are therefore well-suited for implementation on autonomous robotic platforms, providing a critical, self-optimizing tool to accelerate the discovery and assessment of extraterrestrial resources.

Download

Share

COinS