Optimal Differential Drag Control of Small Satellite Constellations

Andrew Blatner

EECS Department
University of California, Berkeley
Technical Report No. UCB/EECS-2018-121
August 10, 2018

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-121.pdf

We analyze and improve two existing techniques for differential drag control of large constellations of small propulsion-less satellites. The system has two subproblems: determining the desired relative ordering of the satellites, referred to as their slotting, and generating an optimal sequence of control inputs to acquire the slotting. One technique, used in production by Earth-imaging company Planet, approximately solves both subproblems with simulated annealing with the objective of maximizing acquired imagery. The other technique, developed academically, ignores slot allocation and generates commands with linear programming, aiming to maximize constellation lifetime. First, we reconcile the practical details of both techniques. Then, we adapt the linear program to new models that better capture Planet's objective. Finally, we develop a fast method for slotting satellites deployed from a single launch. Though we find only small improvements in slot allocation, we provide assurances of optimality and significant improvements in solver time for command generation.

Advisor: Murat Arcak


BibTeX citation:

@mastersthesis{Blatner:EECS-2018-121,
    Author = {Blatner, Andrew},
    Title = {Optimal Differential Drag Control of Small Satellite Constellations},
    School = {EECS Department, University of California, Berkeley},
    Year = {2018},
    Month = {Aug},
    URL = {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-121.html},
    Number = {UCB/EECS-2018-121},
    Abstract = {We analyze and improve two existing techniques for differential drag control of large constellations of small propulsion-less satellites. The system has two subproblems: determining the desired relative ordering of the satellites, referred to as their slotting, and generating an optimal sequence of control inputs to acquire the slotting. One technique, used in production by Earth-imaging company Planet, approximately solves both subproblems with simulated annealing with the objective of maximizing acquired imagery. The other technique, developed academically, ignores slot allocation and generates commands with linear programming, aiming to maximize constellation lifetime. First, we reconcile the practical details of both techniques. Then, we adapt the linear program to new models that better capture Planet's objective. Finally, we develop a fast method for slotting satellites deployed from a single launch. Though we find only small improvements in slot allocation, we provide assurances of optimality and significant improvements in solver time for command generation.}
}

EndNote citation:

%0 Thesis
%A Blatner, Andrew
%T Optimal Differential Drag Control of Small Satellite Constellations
%I EECS Department, University of California, Berkeley
%D 2018
%8 August 10
%@ UCB/EECS-2018-121
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-121.html
%F Blatner:EECS-2018-121