EuroGNC 2019 Paper Abstract

Close

Paper ThA3.3

Biggs, James (Politecnico di Milano), Hugo, Fournier (Politecnico di Milano), Ceccherini, Simone (Politecnico di Milano), Topputo, Francesco (Politecnico di Milano)

Optimal de-tumbling of spacecraft with four Thrusters

Scheduled for presentation during the Regular Session "Spacecraft dynamics and control 2" (ThA3), Thursday, April 4, 2019, 15:00−15:30, BL281.2

5th CEAS Conference on Guidance, Navigation and Control, April 3-5, 2019, Milano, Italy

This information is tentative and subject to change. Compiled on March 29, 2024

Keywords Spacecraft attitude control, Nonlinear control, Optimal control

Abstract

Motivated by a drive towards spacecraft miniaturisation and the desire to undertake more complex missions in deep-space, this paper tackles the problem of de-tumbling spacecraft using a minimal number of attitude thrusters. The control problem addressed is to drive high tip-off angular velocity rates that result from imperfect orbit injection to within a required tolerance using only the on-off switching of 4 thrusters. This paper presents three possible control solutions to this problem (i) a logic-based controller that is simple to implement and requires no tuning (ii) a projective control that aims to replicate an ideal continuous control as closely as possible with the available torques and (iii) a Neural-network-based Predictive Control (NNPC) that is adapted to nonlinear control systems with boolean inputs. The NNPC is based on a Recurrent Neural Network (RNN) using a Nonlinear AutoRegressive exogenous configuration for time propagation of the state in a finite-time horizon optimization. Commonly, for continuous systems, a back-propagation algorithm for the receding horizon optimization is used, but this is not applicable to systems with discrete inputs and so is replaced by a genetic algorithm. In addition a Multi-Layer Perceptron (MLP) is trained off-line with optimal control data obtained with the NNPC resulting in an optimal control that can be implemented on-line with a significantly reduced on-board computational cost. The NNPC performance is compared to the proposed logic-based and projective de-tumbling control laws in simulation of a 12 U CubeSat and is shown to be the most efficient in terms of total impulse requirement.

 

 

All Content © PaperCept, Inc.

This site is protected by copyright and trademark laws under US and International law.
All rights reserved. © 2002-2024 PaperCept, Inc.
Page generated 2024-03-29  00:55:32 PST  Terms of use