Thesis

Genetic programming for intelligent control : real-time guidance for access to space

Creator
Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2024
Thesis identifier
  • T16924
Person Identifier (Local)
  • 201869527
Qualification Level
Qualification Name
Department, School or Faculty
Abstract
  • Reusable Launch Vehicles (RLVs) are becoming a relevant topic within the space industry. These vehicles operate in a broad range of flight envelopes necessitating greater autonomy to compensate for uncertainties or disturbances in real time. Therefore, intelligent Guidance and Control (G&C) architectures are required. The research presented in this thesis aims at investigating the application of Genetic Programming (GP) in an Intelligent Control (IC) setting to perform the real-time guidance of a RLV. The thesis begins with a literature review of the state-of-the-art of G&C and IC methods and applications. Following this, a novel taxonomy of IC was developed, aimed at classifying the applications’ levels of intelligence. The applicability of GP in an IC setting is then investigated, both as a standalone approach and when hybridized with a Neural Network (NN). -- The standalone GP is applied to perform the real-time guidance of a Goddard rocket. This application, the first of its kind for an aerospace vehicle, showcases the ability of GP to produce online guidance commands to successfully track a reference trajectory when external disturbances are applied. Simultaneously, a novel hybrid IC scheme named Genetically Adapted Neural Network-based Intelligent Controller (GANNIC) is introduced. GANNIC is composed of a NN, used as a nonlinear agent, whose weights are updated online by a set of differential equations found offline using GP. Applied to real-time reentry guidance of an RLV, GANNIC proves effective in generating online guidance commands to compensate for uncertainties, and its robustness is assessed through a statistical study. Lastly, as a byproduct of the research, the Inclusive Genetic Programming (IGP), a novel GP heuristic, is introduced. The IGP advances the state-of-the-art of GP algorithms by addressing the population’s diversity issue while demonstrating its efficacy in G&C applications.
Advisor / supervisor
  • Minisci, Edmondo
Resource Type
DOI

关系

项目