Dissertation/Thesis Abstract

Design of a Multi-objective Landing Trajectory Using Artificial Neural Networks
by Dedman, Phoebe Elizabeth, M.S., California State University, Long Beach, 2018, 118; 10837984
Abstract (Summary)

During approach and landing, the HL-20 follows a typical reusable launch vehicle (RLV) autoland trajectory: deep descent, followed by a parabolic flare, and final descent. The trajectory shape is determined by six independent parameters. An artificial neural network (ANN) is designed to generate the trajectory parameters for the HL-20 based on desired objectives using MATLAB®’s Neural Network Toolbox. This research examines three mission objectives: specifying flight time, specifying the final downrange position error, and specifying the average error between the desired angle of attack and actual angle of attack. The ANN successfully produces parameters that meet mission objectives and, in some cases, improve upon nominal errors. It is also demonstrated that the ANN structure and ANN training vectors have a profound impact on the success of the neural network.

Indexing (document details)
Advisor: Shankar, Praveen
Commitee: Gao, Qingbin, Marayong, Panadda
School: California State University, Long Beach
Department: Mechanical and Aerospace Engineering
School Location: United States -- California
Source: MAI 58/02M(E), Masters Abstracts International
Source Type: DISSERTATION
Subjects: Aerospace engineering
Keywords: Neural network, Trajectory
Publication Number: 10837984
ISBN: 9780438387881
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest