Title:
Costates Feedback Control for Mass-Optimal Low-Thrust Transfers

Thumbnail Image
Author(s)
Shimane, Yuri
Izzo, Dario
Ho, Koki
Authors
Person
Advisor(s)
Advisor(s)
Editor(s)
Associated Organization(s)
Organizational Unit
Daniel Guggenheim School of Aerospace Engineering
The Daniel Guggenheim School of Aeronautics was established in 1931, with a name change in 1962 to the School of Aerospace Engineering
Series
Supplementary to
Abstract
Designing efficient low-thrust trajectories involves solving optimal control problems, which can be computationally intensive. One promising approach to tackle this challenge is to use a neural network (NN) as a scheme for obtaining the control input that guides the spacecraft to its targeted orbit. This work explores the use of a NN for learning the costates from the current and targeted states, which can be used together with Pontryagin's maximum principle and optimal control theory to derive the control to provide a feedback loop for controlling the spacecraft. In effect, this is a policy approximation scheme, even though it does not explicitly have the controls as its output. The proposed method is applied to a set of orbits departing from near-ecliptic near-Earth object targeting the Earth's orbit for a mass optimal orbit transfer.
Sponsor
Date Issued
2023-08
Extent
Resource Type
Text
Resource Subtype
Paper
Rights Statement
Unless otherwise noted, all materials are protected under U.S. Copyright Law and all rights are reserved