A Feasibility Study of a Differential Drag Controller for Propulsionless LEO CubeSats

This thesis investigates the feasibility of a novel sigmoid-based differential drag controller for propulsionless LEO CubeSats. The concept of using differential drag to adhere to pre-computed flight plans amid uncertainties in atmospheric density predictions is explored. Solar and geomagnetic impacts on drag authority are often overlooked in passive orbit control methods; this serves as a prime motivator for this research. NASA's GMAT R2022a leverages the UNSW Canberra Space M2 Mission architecture as a high-fidelity simulation platform to study the feasibility of the novel sigmoid differential drag controller. The GMAT simulations incorporate the EGM96 gravity model, the MSISE-90 atmosphere model extended with a CSSI historical space weather file, and a leader-follower twin CubeSat configuration. The leader satellite follows a constant solar/geomagnetic index-derived flight plan, which serves as the reference orbit that the follower satellite tracks via drag modulation. The follower satellite is commanded to propagate through realistic atmospheric density variations, influenced by the solar activity cycle. Three technical chapters systematically evaluate the controller performance and quantitatively assess the drag control authority bounds across varying space weather conditions, satellite configurations, and real reconstructive GPS ephemeris tracking. The first analysis studies the differential drag control response to atmospheric density variations across varying solar activity phases. The controller is tested under periods of quite solar minimum (2018-2019, F10.7 = 67-69 SFU, Kp ~1), unrated transitional phase space weather conditions (2022-2023, F10.7 = 90-150 SFU, Kp ~2), and a G5-rated geomagnetic storm (2024 May Gannon storm, F10.7_max = 237.7 SFU, Kp_max = 8.3). Initialised at a 530 km altitude with an effective nominal M2 surface area of 0.070 m^2, the observed-to-predicted atmospheric density ratios remained within the controller's 57.14%-143.38% drag authority threshold for 82.82% to 97.02% of the quiet periods. During the unrated space weather periods, the controller had varying success in drag authority: 22.08% in 2022 and 94.92% in 2023. However, underpredictions in the F10.7 and Kp indices during storm time led to the controller persistently exceeding its authority bounds, resulting in thousands of kilometres of deviation. The second suite of analyses tests the drag authority limitation at varying LEO altitudes (450-550 km) and area-to-mass ratios (A_min = 0.043 m^2, A_max = 0.2937 m^2). The control authority drops from ~33% to ~17% as altitude increases from 450 km to 550 km due to exponential density decay. Higher area-to-mass ratios increased the usable density range but decreased in-track control authority, whereas lower area-to-mass ratios reduced drag sensitivity and made the satellite less susceptible to environmental perturbations. Additionally, prolonged operation of the differential drag controller confirmed that underpredicted solar/geomagnetic indices, along with outdated flight plans, lead to cumulative orbit drift. The drag control authority is limited by the quality of the input indices used to initialise the controller. The third technical chapter bridged the gap between simulations and flight implementation by propagating the simulated spacecraft to follow a flight plan reconstructed from real M2 GPS ephemeris data. A two-stage orbit determination pipeline is established by using a short-arc batch least squares method followed by an Extended Kalman Filter-Rauch-Tung-Striebel smoother to transform a 66.8-day M2 GPS ephemeris into a reference trajectory. The simulated M2 spacecraft was flown to this GPS-derived flight plan under realistic atmospheric density variations. The spacecraft experienced radial, in-track, and cross-track reference-orbit deviations of 0.3492 km, 6.957 km, and 0.2947 km, respectively. The closed-loop simulation of the differential drag controller using GPS data-driven flight plans validated its potential for implementation as a passive orbit control method. Key findings affirm that flight plan tracking via the sigmoid differential drag controller is feasible for a LEO CubeSat system under varying atmospheric densities, provided the density ratio remains within the control authority bounds. The methodology offers a replicable, non-proprietary framework for testing differential drag control orbit adherence. This research enables precise orbit tracking and improves predictability in the presence of space weather uncertainties, without relying on active propulsion systems. Future work includes the development of an on-board flight plan replanning framework, a hybrid differential drag-lift architecture, and expansion into satellite constellation networks.